This week, the Wall Street Journal published a report detailing developments in China’s efforts to develop advanced chips that may one day allow China to wean its dependence on ASML and Nvidia. Xi confidant Ding Xuexiang heads a so-called “Chip Committee” to oversee the effort, which got off to a rocky start. Last year, Huawei released its Ascend 950, which showed promise. China is requiring companies to apply for waivers to require purchases of Nvidia chips even after Jensen Huang ate Beijing Street food on a charm offensive during Trump’s recent visit to Beijing. Still, the gaps are substantial: Nvidia’s top AI chips still boast around four times the computing power of Huawei’s best Ascend offering.1

China’s push to make its own advanced chips is usually described as a national project. But it is being built in specific places, by different kinds of organizations, under arrangements that vary enormously. China’s indigenous semiconductor push is occurring primarily in three places: the Shenzhen–Dongguan corridor, Shanghai’s Zhangjiang Science city, and Beijing’s Yizhuang Economic and Technological Development Zone. Additionally, inland Wuhan and Hefei are centers for China’s burgeoning memory chip industry, Hefei’s fortunes boosted after the recent IPO of CXMT made it China’s most valuable listed company.

View interactive Map of China’s semiconductor production bases

As Chris Miller has detailed in Chip War, the global semiconductor industry is dispersed across various companies around the world—but in key processes and technologies the ecosystem is highly dependent on one or two companies. ASML in the Netherlands has a virtual monopoly on the production of advanced EUV machines needed to make the most advanced chips. Taiwan-based TSMC is the world’s leading foundry—the production base where chips are made. China, increasingly shut off of the world’s leading chips by U.S. export controls, is trying to get around the restrictions by doing what it always does—doubling down on an all-out state-led effort involving central funding, local competition, and investment in basic research to assist the capital investment of individual companies.

This post explores the places, companies, and ecosystems where China’s indigenous chip effort are happening.

Where China’s Domestic Chip Industry Is Taking Shape

China’s effort to build an independent semiconductor industry is often described as a single national project. On the ground, however, it is taking shape through several distinct regional ecosystems—each with its own companies, government investors, research institutions, and industrial strategy.

The most important clusters are emerging in five places:

• Shenzhen and Dongguan, where Huawei sits at the center of a closely integrated network of fabrication plants, semiconductor-equipment companies, computing infrastructure, and scientific facilities.

• Shanghai’s Zhangjiang Science City, China’s most complete commercial semiconductor ecosystem, spanning chip design, foundries, manufacturing equipment, materials, and capital markets.

• Beijing’s Yizhuang development zone, a major base for semiconductor equipment and mature-node manufacturing, supported by Beijing’s municipal investment institutions. China Academy of Sciences and Tsinghua University are an anchor for research in the city’s northwest, home to new AI chip designer Cambricon

• Wuhan, where the state-backed National Memory Base and YMTC anchor China’s NAND flash industry.

• Hefei, whose locally cultivated CXMT ecosystem has become China’s principal center for DRAM production.

Together, these places reveal how China’s semiconductor campaign actually works. Central policy supplies strategic direction, but municipal governments assemble much of the industry: they provide land, finance fabrication plants, create investment vehicles, establish research facilities, and recruit companies into geographically concentrated supply chains.

The resulting model differs from one city to another.

Shanghai has the deepest market ecosystem, with competing chip designers and publicly listed equipment companies—but its advanced designers still depend on scarce manufacturing capacity at SMIC. Beijing is building on decades of state-owned electronics manufacturing and a strong network of equipment companies and Chinese Academy of Sciences institutes. Wuhan pursued a top-down national megaproject centered on NAND memory. Hefei used patient municipal capital and industrial partnerships to cultivate a DRAM producer after losing the competition to host the original national memory base.

The Shenzhen–Dongguan corridor represents perhaps the most unusual model. Huawei provides technical direction and demand, while state-owned investment platforms hold many of the facilities and companies involved. This arrangement has helped create a network of legally distinct but physically proximate fabrication plants, equipment suppliers, laboratories, and computing centers. It is both an industrial cluster and a large-scale testing ground for domestically produced semiconductor technology.

If you want to know who runs SiCarrier 新凯来, the semiconductor equipment company that has become central to China’s attempt to build its own chip tools, the Chinese corporate registry is not much help. It will tell you the firm is wholly owned by Shenzhen Major Industry Investment Group, a municipal-owned investment platform. For the answer most people in the industry believe, you have to consult the U.S. Department of Commerce, which lists among SiCarrier’s aliases as: Huawei Starlight Engineering Department. The Financial Times reported that Huawei is believed to operate one of them for 7nm production — Kirin mobile chips and Ascend AI processors — with SiCarrier and the memory startup SwaySure running the others.

China has not eliminated its dependence on foreign semiconductor technology. It remains constrained by lithography, advanced manufacturing capacity, specialized materials, and the performance gap between its leading AI chips and those available internationally. But looking only at national funding totals—or at individual champions such as Huawei and SMIC—misses a crucial part of the story.

China is constructing not merely a collection of semiconductor companies, but several geographically rooted production systems. The rest of this report examines who is building them, how they are financed, and what each region reveals about China’s attempt to reproduce one of the world’s most complex industrial ecosystems.

View interactive graphic of China’s semiconductor ownership

For paid subscribers: A detailed tour of the five regional clusters driving China’s semiconductor campaign, including:

• the relationship among Huawei, SiCarrier, and Shenzhen’s municipal investment system;

• how Shanghai combines chip design, fabrication, equipment, research, and capital markets;

• Beijing’s state-backed semiconductor-equipment network;

• the contrasting paths Wuhan and Hefei took into memory-chip production; and

• the research campuses, fabs, government funds, and corporate structures tying these ecosystems together.

Read more

The “Low Trust-Open Source” Paradox of Ai Adoption in China

My research has focused mostly on the infrastructure underpinning AI—data centers, energy, and the application of AI in governance (smart cities) and other infrastructural sectors. Over the last month in China, I had the chance to talk to a number of AI founders and those in China’s cloud companies and public sectors. One of the themes that kept coming up was how lack of trust in data security hinders cloud enterprise adoption, as well as AI enterprise use. This is hard to square with China’s headlong embrace of all things AI. Since the emergence of Deep Seek in early 2025 as an open source rival to leading US AI models, China is sometimes seen to be following a different path that favors “open source” rather than the proprietary enterprise subscription models of Google’s Gemini, Claude, and OpenAi. By many measures, China’s open source models (DeepSeek, Kimi, Moonshot, Zhipu) are more widely used around the world, and even increasingly in the U.S.1 But from a commercial standpoint, AI business models are more mature and already generating far more revenue for U.S. based cloud companies than their Chinese platform counterparts.

The distinction between the U.S. and Chinese AI paths is real. But, a lot of recent commentary on China’s AI ecosystem suggests that China’s “open source” approach arises out of farsighted government promotion of industry adoption and public benefit or high-minded virtues of openness that supposedly contrasts with a proprietary high-fence approach of leading Western cloud and AI platforms. However this narrative greatly oversimplifies actual business dynamics in China. For one, while Deep Seek itself may have developed a culture of research and open source, China’s broader business ecosystem is characterized by mistrust and lack of confidence in data security and data property rights. 2 A central paradox is that China’s embrace of Open Source AI is directly related to a deep-seated and pervasive lack of trust in data security of cloud and AI cloud platforms.

By many measures, China is seeing massive embrace of AI. Nearly all professionals I spoke to were using AI, many using American models via VPN. Airports are blanketed with floor to ceiling ads of leading AI companies, stores sell Ai-enabled glasses (Rokid), AI-enabled pets, and AI-enabled (you name it). When talking about enterprise policy at their companies, nearly people I spoke with said their companies generally prohibit putting any private corporate data or operational data into cloud AI systems—either American or domestic. The preferred approach is to deploy internal AI systems “on premises” neibu bushu 内部部署., which are seen to be more secure. I spoke with the founder of an AI company focused on medical imaging technology (infared) in Shanghai whose cofounder had apparently bought Nvidia chips many years ago and had set up server racks in the company’s headquarters. Many also continue to use VPNs to access Chat GPT and Claude as they are still seen to be superior to Chinese models especially for complex tasks and work.

The divergence between U.S. and Chinese AI ecosystems starts with the basic fact that enterprise cloud adoption in China lags the U.S by a wide margin both for demand and supply deficits, as JS Tan has analyzed here.3 The reasons for this are multifaceted. Chinese companies have lower willingness to pay for recurring subscription-based SaaS services, as is common in many sectors of U.S. business landscape. Sure, China has several leading cloud providers: Ali Cloud, Tencent, Baidu, as well as state-owned offerings from China Telecom and China Mobile that are mostly geared towards government agencies and state-owned enterprises. But another significant reason has to do with a lack of trust, and a lack of enforcement (or a lack of faith in that enforcement) of private property rights, which extends to data ownership rights.

A Shanghai-based employee of an AI startup that is implementing AI in monitoring operations for their company said their managers prevent them from putting any corporate data directly on AI cloud platforms—of either U.S. or Chinese cloud platforms. Employees of larger corporates are generally encouraged to use AI models and cloud platforms within the corporate group. For example, employees of Alibaba use Alicloud, Xiaomi relying on Kingsoft Cloud (owned by Xiaomi Group), Baidu companies use Baidu cloud or Meituan using its own internal model. Many also mentioned the strengths and limitations of some of these models. Meituan’s model is widely seen to be inferior to others but they are encouraged to use it for their own internal operations. The “open source” approach to AI LLM adoption is paradoxically driven by a preference for internally closed data systems.

While trust is one motivation for open source, cost is probably the most important reason. “I think people open source simply because of [neijuan] involution,” my friend working for a cloud company told me. “They cut price for that too, just look at Deep seek and Xiaomi Mimo—Both aimed at gaining market share for longer term gain. Then of course it just so happens that these open source models can be easily deployed privately in private environments.”

Challenge of Data Integration for AI

Discussions of China’s unique approach to AI tend to suggest China is laser-focused on diffusion and enterprise adoption. In 2025, it released the AI+ Policy (an AI update to its 2015 “Internet Plus” Vision), that calls for embedding artificial intelligence across every layer of the economy and society to drive a new era of productivity. Indeed China’s policymakers are focused on diffusing AI and leveraging its benefits for broader economic growth across many industries. But there is a continued challenge of data integration that stems from the same lack of trust in data privacy and property rights that has hindered broader cloud enterprise adoption. Over the past few years, national agencies have unveiled a series of initiatives to promote “data circulation” shuju liutong, the elimination of “data islands” shuju qundao, and the focus on data as a key “new factor of production” or new quality productive forces. Whatever term used, the idea was that data needs to be aggregated and shared in order to realize its full value. In 2023, the National Data Administration was set up, and several cities (including Shanghai, Hangzhou, Suzhou, and Shenzhen) have set up data exchanges where companies are supposed to be able to trade and exchange data. Some of these have focused on data sets needed for Ai model training. But almost all people I spoke with laughed at these policy experiments. “The government puts out policies for things that are the opposite of the reality on the ground. It’s a vision, not reality.” Most of these government or city-backed exchanges see paltry volumes of data actually traded on their “Exchanges.”4 Many people in the industry I spoke to note how companies are hesitant to put their valuable data sets on exchanges, fearing government regulation or theft by competitors. Most companies prefer to buy proprietary training data sets, which as has been shown, often relies on exploitative low-wage human data-labelling enterprises.5

But beyond these mostly political data trading showcases, there is a deeper question of how a lack of trust in data security or data rights actually impacts the development of AI in China. Nothing is stopping Chinese companies from deploying open source models in on-prem servers to help optimize their operations or perform complex tasks. But the lack of serious cloud adoption could hinder some of the economies of scale that come from the aggregation of huge amounts of data on hyperscale cloud platforms. Arguably its the U.S. hyperscalers who stand to benefit the most from AI adoption in the West as companies already using the cloud add on advanced AI tools and services to their existing subscriptions. Cloud companies benefit from lock-in and network effects, where cost of switching providers gives incumbents significant power to lock in existing customers and sell them new Ai services. In China, we do see some uptick in enterprise AI adoption. But the path still seems to favor a more piecemeal approach, where keeping data within secured “data lakes” (either within government agencies or within companies) is paramount. This actually directly contradicts the vision from national agencies and regulators for “free flowing data factors” and aggregation seen in recent initiatives from the 2023 Plan for Building a Digital China to 2025 AI+.

Security first, pay later: SOE/Government procurement

At a state-owned municipal engineering institute I visited in Shanghai, the head of a unit that previously focused on underground tunnel engineering is developing a platform for agentic infrastructure for smart cities, hoping to deploy the service to help automate workflows within city infrastructural management, something long dreamed of in smart city projects. He acknowledged how many of these smart city projects had created pretty but useless dashboards with no function. But with the rise of agents, or zhineng ti 智能体, such data dashboards would not have to rely on one person monitoring information but rather multiple agents processing data, dispatching services, and making decisions. The firm was developing a system for using this to manage trash collection from waste bins across all of Shanghai.

State-owned enterprises are even more wary of putting their data and or corporate processes into proprietary AI cloud environments, and are being encouraged to work with cloud services of major telecoms like China Unicom, China Telecom and China Mobile. But these platforms are widely seen to be lagging behind the private cloud players in terms of technical capacity and sophistication. Huawei, while not formally an SOE, is basically seen to be on the “national team” due to its close military and Party associations. Huawei is thus poised to deploy its AI ecosystem for the public sector and SOE clients. For state-owned enterprises, the focus on data security is even more front of mind. As a friend put it, “SOEs have natural tendency for higher security requirements and doesn’t care that much about cost - they don’t think economically, more politically.”

The relationship of state or government contracts (procurement) for AI enterprise adoption presents an interesting tension. A friend in the industry in Shanghai told me that Sensetime 商汤, which is a partially state-owned Shanghai/Hong Kong based and HK- listed firm in computer vision and facial recognition tech has been declining even as they were described quite recently as “the world’s most valuable AI startup.”6 Their model is not keeping up, and taking on government contracts has been bogging them down. Their facial recognition tech is widely desired for China’s police and national surveillance apparatus. But there is a widespread feeling among industry professionals that government contracts can be both a blessing and a curse for companies. An AI founder in Shanghai told me bluntly “governments are the worst clients, they never pay on time and they keep asking you for more resources and more work.” This echoes much of the research I have done on the exit of private platform firms from the Smart Cities sector. For example, Alibaba helped develop Hangzhou’s City Brain platform which became a model for China’s smart cities nationwide—but the firm laid off most of its smart cities staff and closed its division a few years ago. As a former Ali cloud engineer told me back in 2024, “the government looks at the salaries of big tech engineers and they scoff—why do we need to pay you that much? Why can you just devote more human resources to the project?” Cash-strapped local governments are eager to get contracts for AI on the cheap, but this can suck resources and energy out of companies while failing to deliver solid revenue streams.

Low-Trust, High innovation or AI Involution?

We don’t yet know how this “low trust open source” approach will play out. Despite the pervasive lack of trust of private property/data rights, China’s AI ecosystem is hardly stagnating. The lack of trust in data security may hinder adoption of cloud and enterprise AI services, but it has not hindered diffusion of AI models for deployment within company processes or “on-prem” services. Still, this paradox has important implications for the political economy of AI in China. If leading cloud companies like Alibaba, Tencent, Baidu, and Huawei do not see a massive shift to cloud as a result of AI adoption, who most stands to benefit from the implementation of AI in China? Generally, cloud companies are able to offer customized AI services, which precludes companies having to build out their own server infrastructure internally, which can be inefficient and costly.

In terms of current revenue US frontier labs are operating at a scale roughly 30× Alibaba's MaaS business and over 1,000× the standalone Chinese labs (see below).

Within China, Alibaba is the clear leader in current AI MaaS revenue, with Bytedance’s Volcano in second. (see below).

The term neijuan “involution” has become widely used to describe China’s hyper-competetive business environment that can spawn innovations but also has led to price wars that bring profit margins down. For example, despite the successes of China’s EV industry, many analysts expect significant consolidation in the number of EV firms in the coming years. However, already we see a firm like iFlytek (科大讯飞）which bundles its AI-powered voice recognition software into a diverse suite of hardware products like AI note taking tablets, voice transcription and translation software gaining a larger revenue than any of China’s LLMs and even larger than former darling SenseTime. I personally bought one of iFlytek’s handy transcription devices for transcribing Chinese-language interviews. They even come with ability to detect dialects—Sichuanese, Cantonese, and even Hebei dialect!

Will the same dynamics of “involution” also extend to the AI industry? In other words, will we see rapid innovation in AI application in China alongside falling profit rates for companies involved? Leading tech firms (Ali, Bytedance, Tencent, Baidu) have strong business moats from their hold on key sectors in China. Tencent has been trying to figure out how to integrate AI into its super app Wechat 微信, which is central to life in China. Alibaba, which has a strong position in e-commerce through Taobao, Tmall, and finance, is infusing AI into its existing services, as Grace Shao has noted.7 If this is the case, then AI will help existing platform giants deepen their hold on key sectors. Meanwhile, there are newer AI-native cloud providers like Bytedance’s Volcano 火山引擎, through which ByteDance offers its Doubao LLMs—including the flagship Doubao 2.1Pro. The suite also features advanced multimodal, video generation models like Seeddance (capable of native 4K video generation and 30-second clips), and audio tools. Alibaba released its “Wukong” 悟空 agentic enterprise Ai platform in March 2026.

Then there is the question of Chinese firms going global to get beyond the fierce competition within China. Many Chinese AI firms want to go global immediately, despite the huge size of the domestic market because margins overseas are higher. An employee at an AI video generation APP in Beijing told me 30% of their revenue is coming from North America, with a popular function that allowed users to turn themselves into memes resulting in a popular “put yourself next to Jesus” meme in the U.S. and Europe. So much for Xi Jinping thought on that one!

In terms of actual revenue, a web-based scraping of recent data from corporate reports and news filings suggest that chips firms and those selling actual products are seeing much more current revenue than any of the foundation model companies, despite the leading model companies having sky high valuations. Its possible that the low trust issues discussed here are not a real hindrance to AI innovation in China. China is implementing AI with a ferocity unrivaled anywhere. But who will capture the value from China’s AI economy? There’s the rub! One possibility is that over time a lack of secure revenue for leading cloud providers and/or AI model providers results in a fierce“neijuan-style” competition that the ongoing R&D and investment needed to sustain model development. Of course, the massive amounts of Capex spending by U.S. cloud and AI firms also raises serious questions about the speculative nature of AI in general and the future profitability of U.S. Ai firms, too. China’s national funds and investment vehicles will be more than happy to sustain investment in AI research over the coming years.

China’s AI landscape has importance differences from the U.S., including a preference for open source and industrial diffusion that may offer real long-term advantages. But its also important to be “clear eyed” about the actually existing dynamics of China’s commercial landscape that shape these differences, and will continue to shape the pace and trajectory of AI adoption in China and indeed, the world.

3

Value Added

The Political Economy of China's Cloud

Last week, I published a primer on China’s cloud sector on ChinaTalk, outlining the key reasons why it lags behind its U.S. counterpart. This piece is the deep dive version…

Read more

a year ago · 46 likes · 6 comments · JS Tan

7

AI Proem

Alibaba vs. Tencent's battle to become the most successful internet to AI company

Hi all, I recently received really valuable feedback from a reader. He said he felt I was writing for an audience that already knew the names and the technology, not for people who want to learn about them. I always try to provide context and nuance, so apologies if that hasn’t been coming through lately…

Read more

3 months ago · 23 likes · 5 comments · Grace Shao

Yes, that’s right YOU. You, the boorish uncultured Barbarian who has not spent years dedicated to the study of our difficult impenetrable language and our 5000+ years of glorious history. This criticism gets deployed in various situations: by critical scholars, in heated exchanges on X, and to shut down opinions someone disagrees with. It implies a more sophisticated interlocutor exists somewhere who really gets it, and who would reach different conclusions. Of course, no one has ever been able to define to me who precisely does FULLY understand China or what constitutes full understanding. Indeed, I don’t think anyone can ever fully understand a country, or any complex topic for that matter. I certainly don’t fully understand my own country. If I had, I may have been less surprised on the morning of November 9, 2016 when I awoke to the news of a Trump victory and fellow teachers and classmates in Cambridge, MA in a politically induced depression lasting for weeks and arguably into the present.

Understanding China is indeed a never-ending quest, not unlike the famous episode from the Chinese epic novel “Journey to the West” (西游记) in which Sun Wukong (the monkey king) boasts to the Buddha of his powers and leaps 108,000 li to the edge of the universe to try to prove that he can escape the Buddha’s grip. He lands at the edge of the world only to find five massive pink pillars reaching into the sky and defiantly urinates on one of them, writing “the great sage equal to heaven was here.” The Buddha reveals that the pillars were really his fingers. Wukong is subsequently imprisoned under a mountain for five hundred years so he could learn patience and humility. I hope those embarking on a study of China don’t have to spend 500 years under a mountain (or in a secret detention facility), but more seriously the story is a metaphor for the humility of the individual in a boundless and infinite universe, an apt suggestion of humility for anyone who claims “expertise” or mastery of a subject.

As most of you know, I’m an American who has dedicated the better part of 17 years to studying, researching, and writing about China. I first set foot in Beijing in the cold February of 2009, just months after the city hosted the Olympics in what was seen as a global “coming out party” --certainly not the queer kind, mind you! The period leading up to the Olympics was a high-water mark of China’s global openness, optimism, and U.S-China relations. It would be three years until a deputy of Chongqing Party Secretary Bo Xilai sought refuge in the U.S. consulate in Chengdu, four years until Xi Jinping ascended to the top leadership and seven more until Trump’s victory disrupted assumptions that America would always support free trade. But even in 2009, the signs were there that a shift was underway. I remember my Chinese teacher in Capital Normal University using the ’08 financial crisis to practice Chinese grammatical constructions at a moment when belief in Western global leadership was shattered by the speculative crisis originating in US sub-prime housing markets. Yinwei (because) quanqiu jingji weiji (the global financial crisis) suoyi.. “The Chinese government is implementing various stimulus measures to support the economy.” Only a few months after I first set foot in the country, Western social media platforms were almost entirely blocked in the summer of ’09 following unrest in Urumqi, Xinjiang by Uighur separatists that left scores dead and the city.

But that semester in Beijing and the years living in China as a Princeton in Asia fellow at a small Beijing nonprofit advocating for the preservation of the city’s hutong neighborhoods, and later as a Fulbright Scholar at a university in Xi’an, I took it all in with the wide-eyed wanderlust of a 21-year old set loose in a foreign land. I traveled across 25 provinces by the old clunky Z and T trains, sleeping in软卧 ruanwo and 硬卧yingwo compartments with 3 or 5 other strangers on journeys of up to 20 hours. On one of these from Beijing to Harbin during Chinese New Year, I shared dumplings with strangers as we watched fireworks crackle above each village that we passed in the night. On another overnighter from Jiayuguan, the westernmost part of the great wall in Gansu province to Urumqi in Xinjiang, I mediated a spirited discussion between a young man and an older gentleman on the merits and downsides of American democracy. All of these experiences gave me a profound curiosity and respect for a country that felt often like a world unto itself—not exceptional or something exotic, but indeed an entire world that operated according to its own rules, its own history and traditions, that you could disappear into and potentially spend years learning about only to find that there was always more and so much that you would never know—a good metaphor for scholarship, learning, and life in general.

Since those early years in my 20s roaming the hutongs of Beijing and sleeping in overnight train bunks, I’ve continued engagement with China in various capacities. In 2017, I helped launch a joint research center between Harvard’s Graduate School of Design and Peking University on sustainable urbanism. The collaboration fell apart in 2018, not for political reasons (that I am aware of), but out of mundane squabbles over financing and money between two deans. But only a few months after I left Beijing for another position in Singapore, Trump launched his first tariff war in March of 2018. The tides of the relationship were turning from engagement to outright hostility and competition. During Covid I watched from afar as the country oscillated between a cover up, draconian lockdowns, and then technocratic management of a disease that proved more challenging in open societies with ingrained ideas about personal freedom. Returning to China in 2023 for the first time since 2019, I expected to be met with hostility or suspicion and braced myself for awkward questions and annoying debates with loud taxi drivers. But for the most part these did not occur. I had many conversations with people, some more nationalist or pro-government and others openly critical of their government, but almost none bore personal hostility to me as an American. In a long 顺风 shunfeng1 drive in Inner Mongolia, a driver told me how much he admired America and despised Russia for its war against Ukraine, something directly contradicting China’s tightly controlled messaging on the conflict. When people did debate with me, they usually approached it as a friendly exchange of ideas, playfully challenging me and rehashing certain talking points, but not with any personal animosity.

In the last few years, though, as an academic who continues to write on China, it’s becoming increasingly difficult to do both critical and engaged scholarship on the PRC and also maintain access and relationships in the country. I saw this firsthand in a recent incident that occurred at a leading academic journal. Many of you have followed my writing on Xiong’an New Area, which I began an interest in during 2020 lockdowns in the U.S. as I followed the city’s construction from afar, and finally visited in the summer of 2023. The project is controversial, mostly because it is a signature initiative of Xi Jinping. It has been largely seen as a failure so far, even though investment continues to flow into the city. Initially trying to work through contacts at Tsinghua University to talk to people involved in its planning, I found myself at an impasse. Xiong’an was the “city that must not be named”— no one wanted to be on the record saying anything about it. I asked an old friend who was a professor at a University in Beijing if he would write me a letter of introduction to colleagues at another university to facilitate an interview. I just needed a letter with a nice official chop on it for introduction purposes. He politely declined, fearing it could cause him problems later. Ten years ago, such a favor might have been routine. But today this sort of thing was impossible. Eventually, through introductions of a 30-something urban planner who had worked in Xiong’an for some time, I was able to talk to various people directly involved in planning and design of the city. Nothing high level. But in reality, talking to mid or entry-level workers was more revealing, as higher officials would mostly not divulge anything of interest, anyways. Throughout my research, I tried to maintain an open mind about the city. My first article on the city, published in Cambridge Journal of Regions Economies and Societies, examines the city as a lens into the shifting “urban ideology” of Xi’s China. While I don’t take a position on the city’s “success or failure” (how could I for a so-called “1000-year project) there are certainly aspects of the city that are problematic and worrying. Not least of which is its construction in a floodplain, and the tens of thousands of villagers whose homes were demolished just so they could be moved into apartments in a half-occupied new district built before the rest of the city was completed—of course not something unique to Xiong’an, as most observers of urbanization in China know. After I posted my article on X, someone remarked that, “Oh I didn’t realize a leading journal was ‘pro-China’ now.” I don’t think he read my article, in which I remarked on the city’s various challenges, otherwise he would have not come to that simplistic assessment. But it indicated how anything written about China gets immediately put into the black/white camp—for or against.

I recently contributed another article on the city to a proposed special issue in a leading journal. Compared to my first article, this one does indeed take a more critical tone, particularly related to the issue of the city’s flooding risk.2 Typically, special issues are organized so that an editor invites contributions and goes through an internal review process before individual authors submit to the journal. I went through this process with the SI editors, and submitted. Not two weeks had passed and I woke up one morning to an email telling me my article was “desk rejected” with a generic explanation that it wasn’t novel or a contribution to scholarship. I thought it a bit strange, since the special issue editors had already reviewed it and deemed it ready to send out for review. Later in the day I found out several other authors in the issue had their papers similarly desk rejected soon after submitting, and one of them was a Chinese national based in Hong Kong with another article about China. The editor in chief of the journal, based at a university in China, provided no comment to me or the other authors after repeated requests for clarification, and as of writing had still refused to respond directly even to the special Issue editors, who are themselves established academics at leading research universities. It seemed highly plausible to me and the guest editors that this was a case of political censorship. The lack of communication from the editor in chief was suspicious, as was the fact that most of the rejected articles addressed China in some way. Of course, proving political bias is notoriously hard in academia. Editors could list all sorts of reasons why an article failed to meet their standards. But the overruling of a special issue editor and the lack of communication spoke volumes. The special issue editors promptly decided to pull the special issue.

The episode indicated to me how hard it is becoming to navigate critical scholarship and the deepening culture of paranoia in Chinese academia. To be sure, there are scholars doing engaged critical work in China, and more often than not by those of Chinese ancestry working outside of the mainland. I also don’t want to suggest that doing such work in China is easy and that mainland Chinese scholars don’t face real risks, because they do. The reality is that for academics based at mainland institutions who are under political pressure and surveillance, it is becoming increasingly impossible to fulfill both the obligations of political allegiance in China with the expectation for academic impartiality that (should) be expected for an editor of a leading international journal.

To be sure, I don’t pretend to full objectivity myself, or think that this is a realistic expectation for anyone. But I have tried as much as possible to be objective in my writing and research on China. Sure, I am an American. I have certain values, perhaps most critical of which is a belief in freedom of expression, inquiry, and open debate. No matter how much postcolonial de-constructivist scholars will tell you that rationality is a Western or colonial construct, research and scientific understanding is impossible without certain basic preconditions, most crucial of which is openness to different ideas and viewpoints, and an assumption that all research is open to critique and reassessment. Sure, Trump’s America has also become hostile to scholarly inquiry and academic research, which should be condemned by scholars in the U.S. and abroad just as censorship in other countries is. However, the constraints on academics in China are still far greater than anything faced by U.S.-based academics— funding cuts and Trumpian threats aside. Anyone suggesting an equivalence between the CCP’s control of freedom of expression and what’s happening in the U.S. needs a healthy dose of reality serum. But again, two wrongs don’t make a right—something that often gets lost in arguments on human rights or other sensitive issues between the two countries—“look at homeless people in the U.S., how dare you criticize what we’re doing in Xinjiang!!”— on and on this tiresome exchange often goes in various forms.

On both sides of the Pacific (and globally) academia is often imbued with cultures of hierarchy and deference that means the situation that I experienced recently usually goes unreported. We can and will submit to a different journal. Often, out of fear for career progression, tenure, or maintaining access, academics are too easy to let these improprieties go unreported and unpublicized. If academics can’t hold themselves to standards of accountability, why should anyone trust our scholarship? No wonder why academia has also lost public trust along with many other institutions.

Despite all of this, I continue to stay engaged in research and writing on China, not out of any high-minded heroism but simply because I continue to find it a fascinating place where I learn something new and have rewarding experiences each time I visit. With reels of instant China content available online, it is tempting to think that the slow publishing schedules of academic research are unsuited to the current media environment. But as scholars of disinformation like Yuval Noah Harari have noted, “more information” doesn’t mean better knowledge, or better decisions.3 Despite our access to near real-time information, we (in the non Chinese speaking world) are generally worse off in terms of our knowledge of China than we were with far less readily information, but when we had more sustained reporting and on the ground journalists and researchers, and when there wasn’t as much disinformation floating around. This is not unique to China, it’s a general condition of overly saturated and polarized information ecosystems.

The “you don’t understand China” critique assumes that genuine understanding is achievable — through time, immersion, language, good faith. I think that’s broadly right. Area expertise matters. So does language, and historical depth, and real engagement with Chinese sources and scholarship. But understanding also requires access. It requires being able to ask questions and get honest answers, interlocutors who aren’t operating under surveillance or reporting obligations, archives that are open and statistics that are actually published. Over the past decade those conditions have been steadily degraded. Archives that opened in the 1990s have been shut again. International scholarly collaboration has been brought under new security frameworks. The Foreign NGO Management Law, the data security regulations, the expanded definition of state secrets — these all constrain what can be known, shared, and published about China in concrete ways. Researchers inside the system can’t fully comply with them and maintain the standards of open scholarship. Researchers outside can’t compensate. I know many China scholars who did their research during the period of openness that I first experienced as an undergraduate. Many of them are no longer doing research in China or on China, either due to difficulty of access, concern for the safety of people they interview, or just difficulty of getting approval from home institutions. I see Chinese-born scholars or current PhD students in the U.S. face similar hurdles doing fieldwork in China, even risking being accused of espionage due to their affiliation with American institutions—so the closing off and paranoia doesn’t only affect white foreigners but also Chinese nationals and those of other countries.

That’s why when I see calls from friends and colleagues about the need for more “China expertise” in the U.S, I can’t help but be skeptical. There was a lot of genuine effort to cultivate that expertise over years and decades of research and exchange. Singapore’s NUS has an almost unofficial blanket ban on scholars doing research in China—not because of any restrictions China has imposed on Singapore, with which it maintains generally friendly relations, but because university administrators are afraid of running afoul of China’s 2017 Data Protection Law, which bans the export of data on Chinese individuals outside the country. Research must be done with local partners, but non-mainland academics and even academics of Chinese ancestry based in Singapore are discouraged from pursuing fieldwork there. This probably says more about risk-averse attitudes of Singaporean bureaucrats than it does about China, but the effect is similar—less understanding, less research on the ground, the decreased scholarly exchange, even between two countries that have mostly friendly relations.

What I find surprising, living in Singapore, is how thoroughly the “ten-foot giant” version of China has come to dominate public conversation here too. Singapore has advantages that the Beltway Blob doesn’t — it’s a majority-ethnic-Chinese city-state with deep ties to the mainland, plenty of people with family connections that provide ground-level texture, a long tradition of watching China closely for practical reasons. It should be capable of holding a more complex picture. And yet the conversations I move through in policy and business circles have tilted toward the dazzle of Chinese technological capability — the drone shows, the factory floors, the speed of deployment. Less discussed: youth unemployment running above twenty percent before Beijing stopped publishing the data, or the graduate students delivering food because the white-collar economy can’t absorb them, or the persistent poverty in interior provinces that large showcase projects are partly designed to draw attention away from.

The gaps were on full display during an event last year Singapore. A professor gave a talk on China’s growing cultural influence around the world. I asked a question of how we should reconcile the extreme pessimism inside China, something I have felt viscerally on my recent trips, with the very real progress in certain technological areas. It’s a feeling that anyone who lived through “the good days” would be able to tell you by comparison, but those who didn’t live in China during that time might not. This might just seem like how it’s always been to someone who arrives now for the first time. After the talk, an older Singaporean in the audience came up to me. “So you’ve been to China recently? You didn’t see the robots, and the EVs?” He asked me almost incredulously. “Of course, I did.” How could I not, recalling my visit to an exhibition of consumer robots in Shanghai last October, where I saw robotic capybaras, massage robots and AI-enabled pendants that read your emotions through body temperature. These sexy images of advanced technology are easily transmissible on social media. They are not fake, China has seen real progress and in some areas is indeed world leading, such as in clean energy and electric vehicle production. But simplistic videos of influencers raving about drone delivery in Shenzhen or night light shows in Chongqing cannot convey the stagnant wages, youth unemployment, empty apartment blocks and bankrupt local governments —something that even just a few days on the ground conveyed to me.

The uncomfortable implication is that some of the people most loudly making the “you don’t understand China” argument are embedded in the same institutional apparatus that has made China harder to understand. The critique often comes from genuine frustration with reductive Western coverage, which is real and understandable. But there’s a circularity to it that’s difficult to ignore. I don’t think Western coverage of China is generally good. A lot of it is historically shallow, prone to outdated mental frameworks, more interested in China as a threat to be assessed than a society to understand. The “China threat” discourse in Washington has its own closed loops. There’s a version of China skepticism that functions more as an ideological stance than an analytical one. The “you don’t understand China” critique lands on legitimate targets often enough that it can’t simply be waved away.

But the deterioration of the relationship has produced a polarization of coverage that damages everyone’s understanding, and not just Western journalists. Incentive structures now punish nuance fairly consistently. Algorithms reward clear tribal allegiances — you’re either a China hawk or a panda hugger — and the audience for a carefully qualified, genuinely uncertain analysis is much smaller than the audience that desires to have its existing view confirmed. Chinese censors perform the same function from the other direction, filtering out complexity that doesn’t fit the preferred picture. The space for analysis that is trying to figure something out, rather than perform a position, has become quite limited.

What this produces is a consistent split. American commentary on China has fractured into two camps that rarely engage each other: one cataloguing China’s technological rise — the EVs, the robotics, the solar manufacturing, the infrastructure — and one focused on structural fragilities, the property debt, the demographics, the youth unemployment. Both are drawing on real things. Neither is the whole picture, and the oscillation between anxiety about Chinese power and anticipation of Chinese decline doesn’t quite add up to proper analysis. These things coexist. Treating one side of the ledger as the whole story is its own kind of failure, and it’s not limited to Washington.

Early last year, I was trying to connect with a professor in Guizhou for my research on digital infrastructure in the Gui’an New Area. We were put in touch through mutual contacts. Before we could arrange to meet, she asked me to send her my passport documents. Initially I didn’t think this was a big deal. I knew Chinese university campuses were now routinely requiring IDs including foreign passports for entry. “Ahh, so I can meet you at the university?” I asked her through Wechat. “No no, just so I can meet you somewhere outside for coffee.

Later she told me that she had previously met a foreign academic without reporting the contact to her superiors, and it had caused her problems. She wasn’t going to let that happen again. Sending over my documents gave her a paper trail she could produce if anyone asked. On one level, a sensible precaution. On another, a fairly stark illustration of what “access” now means in practice —bureaucratic precautions required around the simple act of two researchers having a conversation. I sent the documents and went.

The morning I arrived in Gui’an, she picked me up in her car at a subway station near her university. We spent nearly the whole day together, and she drove me out to villages in the surrounding countryside, talking openly about where government programs had worked and where they hadn’t, the effects and limits of investment in data centers on the local economy. We met a young recent graduate working in a village as a new cadre, trying to promote local cultural tourism, who discussed the limits of digitization and a need to focus on cultural value over trendy things like “smart cities.” In the evening she took me to an authentic Guizhou hotpot restaurant near campus and we talked for another couple of hours about her research and her students and her sense of where things were heading. She was warm and generous and entirely without the guardedness that the earlier formalities might have suggested.

The bureaucratic reticence about meeting me were hard to square with the warm hospitality and time she generously shared with me. That gap is part of what makes China genuinely difficult but also rewarding to engage with. It’s not that understanding is impossible. The people I’ve met in China — researchers, planners, students, the professor in Guizhou who spent a day showing me her work are still as curious and eager as before to have engagements and welcome visitors. The difficulty is with the system that now imposes more costs on these types of informal interactions. It’s not impossible to do research in China these days. But you better have a good understanding of what you can and cannot see, and be prepared to accept meeting people outside their places of work, in coffee shops, even in online meeting rooms or in restaurants.

I know people who refuse to go back to China after spending years in the country. Besides the political issues and barriers to access, navigating China’s digital ecosystem of apps can be overwhelming. Luckily I maintained a Chinese bank account from my time working in China, and can easily get a local phone number which allows for more seamless re-entry each time I visit. Without either of those, getting money in and out and paying for things can be a challenge—although much easier since about 2022 when Wechat and Alipay both enabled foreign credit cards to be used for purchases on these two dominant payment platforms. Despite the barriers I would encourage these people to go back. Yes, it is harder to do certain kinds of research in China today. But given the ongoing tension between China and the West, it’s arguably more important than ever before for Americans and others in particular to have more nuanced and layered understanding of the country, its successes, challenges, and other aspects of life that are no longer deemed as important to outside observers as AI and robots—in other words, everyday life. Learning and knowledge is not a finite process, it is never over. Just like Wukong, you could find yourself reaching the limits of available knowledge only to realize you never left the palm of the Buddha’s hand in the first place.

On a recent visit to Thailand’s industrial heartland of Chongburi, Somnuck Jongmeewasin, research director at EEC Watch and a process engineer by training, describes how the construction of a new hyperscale data center is damaging the delicate coastal mangrove forest of Khlong Tamru in Chonburi.

“The water temperature is affected by the release of water from the DC cooling,” he told me. “This can kill off sea life in the area quickly. I am a former engineer — you can’t fool me with these things.”

For ten years, Somnuck has brought attention to the environmental problems of the Eastern Economic Corridor, a flagship initiative of Thailand’s Junta government under Prayuth Chan-Ocha to spur new investment into Thailand’s industrial heartland, which was originally developed around Japanese infrastructure assistance and auto investment in the 1980s. In the early 1980s, the Japanese government — through JICA and a constellation of private automotive and petrochemical firms — financed and designed the Eastern Seaboard Development Programme, one of the largest aid-funded industrial infrastructure projects in Southeast Asian history. The design followed a deliberate two-zone spatial logic: heavy petrochemical industry concentrated at Map Ta Phut in Rayong; automotive and electronics assembly anchored in Chonburi at the Amata City and later Amata Nakorn estates. Water infrastructure, power distribution, road connections, and the deep-water port at Laem Chabang were engineered specifically to support these two zones and their resource needs.

The Eastern Seaboard became the heart of Thailand’s automotive export industry, hosting Toyota, Honda, Isuzu, Mitsubishi, and a deep Japanese-dominated supply chain. By the 2000s, Thailand was producing over a million vehicles annually — nearly all assembled in or near what is now the EEC. The infrastructure built for that industry persists as a fixed spatial inheritance. Reliable high-voltage power, a deep-water container port, industrial-grade water supply, and flat, large-footprint serviced land are what automotive manufacturing requires. They are also, with minimal modification, ideal for what EV assembly plants and hyperscale data centers require. Chinese capital arriving in the EEC is not creating infrastructure from scratch but rather layering onto an existing infrastructural landscape.

The Prayut government’s plan for the EEC called for seven new tech clusters, a high-speed rail link with Bangkok and its airports, and a new aerotropolis at U-Tapao in Rayong. None of these signature initiatives have born much fruit so far. But in recent years, the FDI from Chinese batteries and EV manufacturing has begun to transform the area.

Click for Interactive Chart

Electric vehicles: occupying the Japanese footprint

The most consequential Chinese investment in the EEC — measured by capital committed, employment, and spatial impact — is in electric vehicles and batteries,. Between 2022 and 2025, at least five Chinese automotive brands secured BOI approval for assembly operations in the EEC, with collective announced capacity exceeding 600,000 units per year. The spatial pattern is telling: nearly all of them have located in the EEC, the area that Japanese automotive investment transformed into the so-called Detroit of the East. So far, the area is not going the route of Detroit in de-industrializing, but rather transforming from the diesel era to the EV era.

Great Wall Motor arrived first. In 2020, as domestic Chinese market competition intensified, GWM acquired a mothballed General Motors assembly plant in Rayong — a facility that GM had built during the high point of Japanese-era automotive expansion and abandoned in the 2010s. GWM converted it for its ORA and Haval electric and hybrid lines, with capacity around 80,000 units per year. The symbolism is pointed: a Chinese OEM inheriting the physical shell of an American plant, on land serviced by Japanese-funded infrastructure, to produce EVs for a Southeast Asian market.

BYD 比亚迪 followed with even greater ambition. Having secured BOI approval in 2022, BYD commenced construction at WHA Eastern Seaboard Industrial Estate 36 in Rayong and opened in 2024, with a designed capacity of 150,000 vehicles per year. The facility is BYD’s first overseas assembly plant in Southeast Asia and is positioned as the export hub for the ASEAN market. Chang’an Automobile 长安汽车 has received BOI approval for a comparable 100,000-unit facility in Rayong. Chery, another Chinese OEM with significant export ambitions, is also BOI-approved for EEC production. SAIC-CP, a joint venture between China’s SAIC Motor and CP Group, has operated MG brand vehicles in Thailand for over a decade through a Thai-Chinese partnership structure that insulates the investment from straightforward classification as either foreign or domestic.

Batteries: the deeper supply chain integration

The more structurally significant dimension of Chinese industrial investment in the EEC is not vehicle assembly but battery and energy storage supply chain. Vehicle assembly can be relocated; battery manufacturing, once established, creates deeper dependencies through materials sourcing, technical workforce, and capital intensity.

CATL（宁德时代） the world’s largest EV battery manufacturer, has structured its Thai entry through a joint venture with Arun Plus, a subsidiary of PTT (Thailand’s state-owned petroleum company). The JV model is significant: it mirrors CATL’s approach in other markets of partnering with established state-linked entities to manage regulatory and political risk, while gaining access to land, grid connections, and import exemptions that independent foreign investment cannot easily obtain. The PTT connection also gives CATL positioning in Thailand’s energy transition — PTT is simultaneously a major gas supplier, a renewable energy investor, and the state institution most directly exposed to the structural shift from fossil fuels to electrification.

EVE Energy, a Shenzhen-listed battery cell manufacturer, received BOI approval for EEC production — part of a broader EVE expansion across Southeast Asia targeting lithium iron phosphate (LFP) cell production for the regional market. Sunwoda 欣旺达, another Shenzhen firm specializing in battery packs and consumer electronics batteries, has similarly received BOI approval for EEC operations. These are not household names in the way BYD and CATL are, but their presence in the EEC indicates that the investment wave is reaching into second-tier supply chain depth — the kind of integration that preceded the establishment of a Japan-centered automobile ecosystem in the 1990s.

Japanese battery and component suppliers followed Japanese automotive OEMs into the Eastern Seaboard across the 1980s and 1990s, creating a supply chain network dense enough to constitute a genuine agglomeration economy — firms could locate in Chonburi or Rayong and source components, tooling, and technical labor locally. Chinese battery supply chain investment is following a similar trajectory, but compressed into a much shorter timeframe and driven partly by overcapacity at home rather than by organic export demand. CATL, BYD, and EVE are all dealing with brutal domestic margin compression as Chinese EV penetration reaches saturation. Thailand — with its BOI incentives, existing industrial estates, and proximity to ASEAN’s growing car market — offers a relief valve.

Thailand as burgeoning ASEAN data center hub

Thailand is now seeing a ramp up in data center investment. In 2025 alone, Thailand’s BOI approved 36 data center projects worth approximately US$23 billion, with roughly two-thirds concentrated in the EEC. The country is targeting 1 gigawatt of data center capacity by 2027, tripling its 2024 baseline.1 On the data center side, investment is more balanced across Chinese, Western, Japanese, and Singaporean capital. AWS, Google, and Microsoft have all made significant commitments and opened dedicated cloud regions in Thailand in the past few years. Amazon’s $5bn commitment over 15 years is the largest in absolute value, although its immediate investments have not been indicated out of this large commitment. Google opened its cloud region in January of 2026, and Microsoft announced similar intentions in early 2026.2

Whereas smaller co-location facilities had previously clustered in and around Bangkok itself, the EEC’s cheaper land, energy, and water availability makes it a natural destination for hyperscale facilities, which require large amounts of all those inputs. Bridge Data Centers, a Chinese-founded company (but with significant backing by Bain Capital), is building a large 100+Mw facility in Chonburi on a coastal mangrove forest described in the article’s opening. DayOne, the rebranded international unit of Chinese data center firm GDS, broke ground on a 1GW campus at Amata City Chonburi in March 2025. Beijing Haoyang Cloud & Data Technology purchased land for a 300MW facility at US$2.16 billion WHA Eastern Seaboard Industrial Estate 4 in Rayong in June 2025, in the same estate where Chang’an motors is building its new plant, and where AWS has already opened one of its new data centers. Galaxy Data Center (owned by Hoyinn Technologies) is planning a 1GW facility in Rayong’s Ban Chang district.

Many foreign cloud firms are partnering with local intermediaries. ST Telemedia, a Singaporean co-location DC operator, has a JV with Frasers, a division of the Thai conglomerate ThaiBev. Google and Microsoft both announced energy partnerships with Gulf Energy to supply their new facilities. True Digital, a subsidiary of the powerful CP Group, has partnerships to offer access to many of the leading cloud platforms through its data centers and has close partnerships with Huawei in particular.

What distinguishes the data center buildout spatially from the EV story is its relationship to submarine cable infrastructure. Laem Chabang Port, built with Japanese ODA in the 1980s as a container terminal, is also where several regional and international submarine cables make landfall. Google’s TalayLink cable (connecting Thailand and Australia, announced alongside its Bangkok cloud region launch in January 2026) lands there, as do older regional cables.3 Data center location in Chonburi and Rayong is determined not only by energy and water availability but by latency and connectivity: these provinces offer large serviced land, industrial-grade power, and fiber connectivity.

Global Supply Chains, Local Environmental Consequences

In one sense, the wave of investment into Thailand’s EEC is not new—it is merely the latest instance Thailand’s engagement with foreign investment in manufacturing. Today’s EEC is being remade by Chinese investment in battery and EVs, while data center investment in attracting a wider range of capital including Chinese, Western, Middle Eastern and Asian DC operators. Chinese industrial presence in the EEC spans the production of physical goods (vehicles, battery cells), digital infrastructure (compute, storage), and energy systems (grid-connected power, renewable PPAs). This is not coordination in any centralized sense — BYD, CATL, Haoyang, and DayOne are operating on independent commercial logics. But the spatial outcome of those independent decisions is a Chinese industrial footprint in the EEC that is wider and deeper than any single category of investment suggests.

The spatial consequences of this largely unregulated investment is also a continuation of earlier logics—Thai industrial estate developers ( WHA, Amata, and Rojana) acquire larger tracts of land and sell or lease to foreign investors. The scale of new facilities is visible in aerial photography of the sites—massive enclose compounds. In the case of BYD’s WHA 36 site in Rayong, their factory includes worker housing, production plants, water treatment facilities, and energy production.

The question of local benefit remains debatable. On the one hand, EV and battery ecosystems could lead to a genuinely dense cluster of related industries and spur network effects, shaping the EEC into an EV industrial base. However, numerous reports suggest much of the labor is imported—on both the higher and lower end of the value chain.

Chinese companies bring in their own labor. For the Bridge data center, its mostly Myanmar and Cambodian workers usually.”
— Somnuck Jongmeewasin, EEC Watch, May 2025

This pattern of labor importation is consistent across EV assembly and data center construction. It reflects a strategic choice to keep costs low and supply chains controlled, but it limits the local employment multiplier that BOI incentives are ostensibly designed to generate. The Thai government’s BOI negotiations with data center investors have run into exactly this tension: the Board reduced corporate tax exemptions for data centers from 10 to 8 years partly in response to arguments that the employment benefits were not commensurate with the fiscal concessions. Data center operators argue that DC investment spurs broader digital transformation, but this may hard to link directly to any one specific project and the benefits will not generally accrue directly in the communities where the facilities locate but rather in the companies that use them—and they are most likely in Bangkok itself, Singapore, or beyond.

The environmental impact of largely unregulated investment could be massive while local benefits fairly limited. The 2018 EEC Act provided legal authority for formally overriding local zoning and planning in the three provinces. The EEC Policy Committee is chaired by Thailand’s Prime Minister, and has authority to grant concessions and oversee development in the area. This, coupled with an almost total absence of environmental assessment, lack of community consultation around major facilities were features of the zone’s designation, not symptoms. They were intended to make the EEC frictionless for foreign capital. Under Japanese developmental investment, this produced agricultural land loss and water competition that farmers in Chonburi and Rayong are still living with. Under the current wave, those same dynamics are being accelerated by data centers competing with households for reservoir water, and by battery factories introducing new chemical handling risks into an already over-stressed industrial landscape.

The dominant frame for discussing AI in Southeast Asia has been geopolitical. This frame is not entirely wrong — U.S. export controls on advanced chips, Chinese firms’ regional infrastructure push, and the bifurcation of the global economy are real. But this view misses a richer picture of the domestic political-economic dynamics driving AI adoption across the region.

Across Malaysia, Indonesia, and Thailand, governments are not simply tilting toward Washington or Beijing. All the major countries in the region are actively promoting AI. They are encouraging foreign investment in cloud, be it from Chinese or American cloud firms. There is also another aspect to the story: the critical role of domestic conglomerates in the energy, land, water.

In each country, a similar pattern is playing out: discourses of “AI sovereignty” are invoked primarily by domestic actors (government and domestic conglomerates) with prior state connections to capture infrastructure and land rents. Pressure to adopt AI quickly and competitive dynamics between countries are leading countries to aggressively court foreign infrastructure investment. Yet, the governance and institutional mechanisms that might convert infrastructure into genuine domestic innovation ecosystems are systematically underfunded or co-opted. This is a story that is not fully accounted for in narratives of U.S-China rivalry.

Despite proclamations of the region’s shift toward China or focus on China’s “Digital Belt and Road,” 1in terms of cloud and AI, American hyperscalers are widely preferred across the region in terms of the capacity, existing corporate relationships, and relationships with governments in the region (see below figure). This has little to do with actions of particular U.S. administrations and everything to do with capabilities, compute power, and existing lock-in effects of enterprise cloud use. True, if public sentiment shifts even further against U.S-based companies (some uproar in Malaysia about Microsoft’s contracts with the Israeli military), government could be forced to shift its spending on U.S. cloud. Chinese firms are present (Alibaba Cloud, Tencent, Huawei) along with co-location operators (Day One, Bridge) but they are working alongside the cloud infrastructure of the American giants, not against it. They also compete in different market segments, targeting lower end basic cloud services.

Tracking Cloud Investment in ASEAN

Access the ASEAN AI investment tracker here.

Alongside direct investment in infrastructure, cloud companies have generally committed to various “upskilling initiatives” in each country as a form of technology transfer. While the specifics or follow-up data on the results from these training programs are hard to come by, they do provide an imperfect metric for company’s various commitments to host country governments. As you can see, upskilling initiatives of host governments are still larger than any of those by cloud hyperscalers, but among the big cloud companies, Microsoft has been the most out front in terms of explicit skilling commitments in the region.

Access the “ASEAN upskilling tracker” I put together here.

The most important axis of tension in Southeast Asia's AI political economy is not East–West. It is between the interests of the states, conglomerates, and hyperscalers who are jointly constructing this infrastructure — and who all benefit from the current arrangement — and the broader populations who will live with its consequences: the energy and water demands on communities near data center clusters, the concentration of AI capability in large enterprise clients rather than SMEs and public services, and the widening gap between the region's participation in AI deployment and its capacity for AI development.

Malaysia: Digital Hub of ASEAN?

Malaysia has been laying the groundwork for digital infrastructure for some time, dating as far back to the Multimedia Supercorridor, a signature policy of Mahatir’s government in the 1990s. But Singapore’s data center moratorium, which was in place from 2019-2023, was the most significant recent catalyst that jumpstarted investment into Malaysia’s cloud infrastructure particularly in Johor, which neighbors Singapore. Since then, Malaysia has led ASEAN in data center investment. The country has aggressively courted foreign investment in cloud. With a stated intention to become ASEAN’s digital hub, Malaysia has big ambitions. But there are key questions over the legal and institutional support necessary to translate infrastructure investment into local innovation capacity. Much of the DC investment in Johor serves Singaporean or Chinese businesses, thus raising questions over whether DC investment will spillover into the domestic economy. However, some Malaysian domestic conglomerates ARE seeing immediate benefits from the AI boom in terms of capturing the increased demand for energy and water infrastructure.

YTL was founded by the Yeoh Tiong Lay in 1956, and is one of Malaysia’s largest Chinese owned conglomerates. With interests in real estate, power, and infrastructure, YTL has relied on state patronage and stable monopolies for its diversified business interests. YTL is using its infrastructural assets to compete in the AI space. In 2025 it completed the YTL Green Power Data Center in Johor, which uses Nvidia’s high-end AI processing chips. Anticipating the crucial role of water for DC cooling, YTL bought out Ranhill, which has an exclusive concession to supply tap water in the state of Johor, in 2024. Meanwhile, YTL is also developing AI applications, such as ILMU—a local LLM trained on Malaysian languages and data and branded as “100% made in Malaysia by Malaysians for Malaysians.”2 Of course, this obscures the reliance on Lamma for open source weights—but ILMU represents a significant push for a local LLM nonetheless.

YTL is also developing Ryt Bank, an “AI-powered bank,” that is a joint venture with Singapore’s Sea corporation. Another Malaysian startup Mesalotica has developed another open source local LLM, trained on local dialects and language.3 On the KLIA train from the airport to Kuala Lumpur City Center, you’re riding on a train operated by YTL, with wifi (intermmitent) access provided through their Yes mobile network, looking at advertisements on the seat backs for Ryt Bank. This is vertical integration of telcom, infrastructure, and AI applications, Southeast Asian style.

There is definite anxiety by some that the influx of investment in cloud is not providing benefits to most Malaysians. However, an official who has worked on cloud computing for one of Malaysia’s digital agencies under the Ministry of Digital told me he thought the fear over Ai sovereignty was misplaced. “I think the more important thing is Malaysia can develop applications using cloud platforms. I don’t think the government or domestic actors will be able to provide the types of sophisticated cloud computing services that the top hyperscalers offer.” He noted that sovereignty discourses are most commonly invoked by local actors like YTL who are trying to obtain government contracts for cloud. Others also recognize that the hyperscalers and YTL are serving distinct niches but are not necessarily at odds with each other . “In general we see these firms playing their own role in Malaysia’s AI ecosystem,” an official with Malaysia’s MyDigital agency tells me. “These services drive broader demand for AI in Malaysia, so I don’t think what they’re doing competes directly with the likes of Microsoft, AWS, and Google.”

Indonesia: large market, barriers to investment

Former President Joko Widodo directed the national government in 2021 to accelerate AI capabilities, declaring: “The world today is in a ‘war’ to gain AI capabilities. The competition to control AI is comparable to the space race during the Cold War.” During the Widodo administration, the country released several high-level policy documents, including the “Making Indonesia 4.0” roadmap, the “Digital Indonesia Roadmap 2021–2024,” and the “National Strategy for Artificial Intelligence 2020–2045” (Strategi Nasional Kecerdasan Artifisial). However, Indonesia has yet to issue any binding AI regulation. The Ministry of Communication’s 2023 AI Circular offered only nonbinding guidelines, and Indonesia’s AI governance remains fragmented across the Coordinating Ministry for Economic Affairs and the Ministry of Communication and Digital Affairs (Komdigi). “In Indonesia, in the early stage, the main player was BPPT. But now Komdigi is the primary actor, followed by the private sector.” Various experts indicated the likelihood of an upcoming presidential regulation on AI, with some preferring a flexible non-binding framework given the rapidly changing nature of the technology.

Cloud companies have begun to invest in Indonesia. In April 2024, Microsoft announced a $1.7 billion investment over four years in cloud and AI infrastructure in Indonesia, the largest in the company’s 29-year history in the country. NVIDIA and Indosat Ooredoo Hutchison followed with a $200 million partnership to construct an AI center in Surakarta in the Solo Technopark4 and billed as the country’s first sovereign AI computing facility “GPU Merdeka.”5 The NVIDIA partnership with Indosat follows a similar approach as in Malaysia: provide advanced AI chips to a local conglomerate to build out a “sovereign AI” data center based in the country and operated by local company. Of course the reliance on NVIDIA chips means its not truly “sovereign,” but it is meant to suggest a move away from relying only on the big cloud hyperscalers (whether American or Chinese) for AI applications and training capacity.

Google, AWS, Tencent Cloud, and domestic operators including DCI Indonesia, EDGE DC, and NTT Indonesia have similarly expanded. Indonesian data center operators still face legal uncertainty and higher electricity costs. Indonesia’s electricity consumption per capita stood at just 1,337 kWh in 2023 — substantially below Vietnam, Thailand, Malaysia, and Singapore — reflecting incomplete industrialization that limits electricity demand and constrains AI infrastructure build-out.

Thailand: catching up in DC investment, uncertain digital adaptation

Thailand has arrived late to the regional data center boom — Singapore's 2019–2022 moratorium and Malaysia's subsequent surge came first — but it is now moving with remarkable speed. Google Cloud's Bangkok region launched in January 2026. Microsoft committed over $1B to an Azure region in March, partnering with Gulf Energy and CP Group. TikTok has pledged $8.8B over five years. Chinese operator Haoyang Data has purchased land in Rayong for a 300 MW hyperscale campus. AWS is live. Alibaba launched its second Thai data center in early 2026. Japanese telecom operator NTT has also announced new investment in Amata City Chonburi, an industrial area with a strong presence of Japanese companies.

The domestic partners in these deals — Gulf Energy, True/CP Group, WHA — follow a familiar pattern: conglomerates with prior energy, real estate, and infrastructure footprints, now repositioning as critical intermediaries in the AI economy. Thai Billionaire Sarath Ratanavadi has ascended the ranks of Thailand’s rich list through his role in Gulf Energy. Gulf became an important partner of Royally owned businesses Siam Cement Group. In 2024, Gulf acquired Thailand’s second largest telecom AIS (originally founded by Thaksin Shinawatra and later sold to Singaporean investors). With his ownership of key energy infrastructure and a major telecom, Gulf is poised to capture rents and value from the growth in the digital economy.Gulf has partnered with both Microsoft and Google Cloud as a local energy and land supplier. However unlike Malaysia’s YTL and Indonesia’s Indosat, Gulf has yet to unveil any significant sovereign AI initiatives or digital applications of their own.

Hyperscale data center expansion has mainly come in the Samut Prakan province around Bangkok’s main airport and the Eastern Seaboard area (Chonburi, Rayong). This area was set up as a special economic zone by the Prayut government in 2017, focusing on attracting Chinese investment in electric vehicles and batteries. The area was originally developed into the country’s largest export-oriented zone through investment of Japanese manufacturers in the 1980s. With significant investments of Chinese firms BYD, CATL, Great Wall Motors, and others, Chinese EV companies are quickly transforming Thailand’s auto sector. In terms of cloud, investment has located in industrial parks near major power stations. The area also benefits from having an industrial water system, which was planned and constructed through Japanese loans over the last few decades.

Thai LLMs

Thai is a notoriously hard language to learn (I tried myself, and failed miserable despite several years of courses). It’s inconsistent alphabet system, tones make it even harder for AI models. One of the first Thai language models Typhoon was developed by SCB 10X, a venture arm under Siam Commercial Bank, one of the country’s major banks and the one owned directly the Royal Family. 6

This post does not examine feng shui as a private or popular practice, although it certainly continues in this aspect, particularly in Hong Kong, southern China, and overseas Chinese communities in Southeast Asia. Instead, here I trace its evolution into a state-led logic of legitimation, visible in large-scale planning and environmental enforcement—from the siting and design of Shenzhen’s Futian CBD, Guangzhou’s Baiyun District, Xi’an’s Qinling villa scandal, and the Xiong’an New Area. Xi Jinping himself is said to be a believer in the power of feng shui. Thus, the concept persists not as mysticism, but as a durable grammar through which the Chinese state organizes space, nature, and power. Far from vanishing under socialism or technocratic planning, feng shui was absorbed and rearticulated, translated into the language of ecology, environmental protection, and spatial order. The malleability and variety of concepts in feng shui lends itself to serve as a convenient vocabulary of legitimation for state projects.

To be sure, the officially sanctioned feng shui is a watered down and simplified version. On the one hand, the state has tried to limit supposedly superstitious practices like feng shui—that contradict the CCP’s official Marxist non-religious position. Many officials targeted in Xi’s endless anti-corruption crusade have been cited as having practiced too much feng shui. Former security minister Zhou Yongkang 周永康, one of Xi’s earliest and most important targets, was alleged to have consulted advice from a qigong master, who was one of those who testified against Zhou in court. In 2021, tourism minister Li Jinzao was sacked, and one of the charges against him included carrying out “superstitious activities.” Former Hebei party secretary Zhou Benshun was also sacked and investigated for “seeking the protection of supernatural beings.”3 On the other hand, Xi himself has made protection and promotion of “excellent Chinese traditional culture” 优秀传统文化 as a form of legitimacy. This party-sanctioned version of feng shui dispences with superstitious aspects of fortune telling and emphasizes basic principles of city and regional planning—orienting buildings and new districts in relation to important mountains or “dragon veins,” and with water to the south, and following grid-like planning. Many traditional Chinese capitals, from Chang’an to Ming Beijing, were laid out according to these principles. New urban districts in recent years have also been planned according to principles of feng shui.

Shenzhen Futian CBD

Shenzhen was China’s first urban experiment of reform and opening up. In 1979, Deng Xiaoping 邓小平 tasked Yuan Geng, a director of Hong-Kong based red enterprise China Merchants to develop a new industrial zone in 蛇口 Shekou, just across the water from Hong Kong. Shekou was an experiment in granting autonomy to capital, allowing limited foreign investment, and even saw short-lived experiments in local newspapers and self governance. Shenzhen’s success was not assured in its early years, and was only concretized after Deng’s famous Southern Tour in 1992 when he visited the city again to signal his support for the city.4 Shenzhen developed through a series of enclaves parceled out to various state-owned enterprises—China Merchants 招商局, China Southern Petroleom 南油, Overseas Chinese Town 华侨城, and various manufacturing SOEs in the Huaqiang Bei 华强北 area. 5 In the 1980s the city began planning a new CBD that would create a more formal city center in what had hitherto been a sprawling decentralized growth pattern across the various zones.

The Futian CBD project created a formal, orthogonal grid of 9 square blocks （echoing the well-field system or the ritual urban plan proscribed in the kao gong ji), aligned to the cardinal directions with a central axis aligned to the hills of Lianhuashan Park. A public open space spine runs along the central axis, underneath a sweeping curved structure that From a platform atop the park, a statue of the city’s patron saint Deng Xiaoping gazes out over the skyline below. Whereas Shenzhen began its growth in a relatively decentralized unplanned way, the Futian CBD was about installing a more formally designed administrative district, symbolizing greater power of the local government over the city, which had formerly been driven by enclaves controlled by various SOEs.

Guangzhou Baiyun District

Guangzhou, the ancient capital of Guangdong province and one of China’s largest cities, is thought to have excellent feng shui. A mountain range runs from the northern part of the province down into the city proper, forming a ridge against which the city is nestled, and to the south of the city is the Pearl River 珠江. The city has since expanded in every direction, including to the south bank of the river. The Baiyun District stretches north-south, paralleling the ridge of the Baiyun mountains, which are being conserved as an ecological spine for the city.

Xiong’an New Area

In 2017, Xiong’an New Area was announced as a new area 100 km southwest of Beijing to relocate “non-capital functions” from the congested crowded Northern Capital. Xiong’an is a signature project of Xi Jinping, and has taken on symbolic importance as a reflection of the changes Xi wishes to see in Chinese urbanization. The city is low-lying, integrated with a waterway system, and strongly aligned along central axes—both East-West main transport corridor, and a central North-South spine of Open Space. The main commercial district (Qidong District) is located to the East of the central axis—just as Beijing’s Guomao is directly East of Tiananmen Square on Chang’an Avenue. Tang Dynasty Chang’an also placed the main markets at equidistant locations from the central axis—the layout reinforce a core dynamic of Confucian hierarchy that the state is superior to commerce, and merchants at the lowest rung of society.

Xiong’an is located directly north of Baiyangdian—so with water to its south this might seem like ideal feng shui. But if we zoom out, the area where the city is being built is extremely low-lying, 8 or 9 meters below the long-term flood line. According to traditional feng shui principles, cities are sited on higher elevation. This is practical as well as spiritual—water flows out of the city, reducing flood risk. So to make Xiong’an safe, an enormous investment in new infrastructures are being built to prevent flooding—a ring levee system designed to prevent 200-yr floods, new sluice gates and pump stations, and new forest parks designed to absorb potential floods.

Xu Kuangdi, former Mayor of Shanghai who in 2014 was appointed to lead the advisory panel for the Coordinated Development of Jing-Jin-Ji, had elsewhere portrayed the choice of location as a simple matter of rational scientific planning also noted that, “traditional Chinese culture has the idea of mountains to position city location. Directly south of Beijing’s central axis is Bazhou, but the geological condition there was not suitable for a new city. The three counties chosen for Xiong’an are relatively sparse in population and lie directly south of Tanzhe Temple. This is an ancient temple with origins as far as back as the Western Jin Dynasty (307 AD). “Before there was Beijing there was Tanzhe Temple 潭柘寺. Relying on this axis, Xiong’an was selected” (Zeng, 2017). 6Aligning the city’s central axis with the temple can be said to endow the city with a degree of protection and stability. Given the controversy surrounding Xiong’an’s low-lying location, principles of feng shui offered a convenient post-facto rationalization.

Xi’an Qinling Villas and National Archives Project

The Xi’an branch of the National Archives of Publications and Culture is located right up against the Qinling mountains, which separate North and South China, and have been a home for Daoist mystics and recluses for centuries. The Xi’an Branch was designed by well-known Xi’an architect Zhang Jinqiu 张锦秋, who has designed numerous cultural buildings in the city such as the Sha’anxi Provincial Museum 陕西省博物馆. The location of the Xi’an branch is interesting—note in the below image the central axis of the archive is aligned with the mountain peak behind it—a so-called “dragon’s vein” or longmai, in the traditions of Feng Shui.

The area is also perhaps not coincidentally where illegal villas were demolished in a 2016 scandal that brought down much of Xi’an and Sha’anxi province’s top leadership. The“Qinling Ecological Environment Protection Regulations” were issued in 2008. However, since 2003, developers have continued to build villas on the northern foothills of the Qinling Mountain. Xi Jinping, when visiting the area remarked, “let this be a warning that cadres must always protect the Qinling mountains.” The location of the archives built nearby these demolished villas, an investigation personally overseen by Xi, is yet another reminder of the power of the central state to safeguard the nation’s environment over the commercial interests of local officials. The construction of the new archives branch, known to be a personal project of Xi’s, was a symbolic statement of the central state’s role in guarding China’s natural environment such as the dragon veins of the Qinling mountains. Xi visited Qinling again in January of this year to emphasize value of ecological conservation and guard against illegal construction.7

3

Beijing Channel

Chinese corrupt officials and superstitious activities

A few days ago another “tiger” was brought down by China’s anti-graft campaign, vice minister of culture and tourism Li Jinzao…

Read more

6 years ago · 1 like · 2 comments · Yang Liu

This is one of the eight data center clusters数据中心集群 of China’s Eastern Data Western Compute dongshu xisuan 东数西算 project. In this post, I’ll touch on some of the dynamics of China’s AI infrastructure and data center development through Zhangjiakou. Why this city of all the eight clusters? Zhangjiakou is adjacent to Beijing, and therefore offers lower latency for cloud//AI usage than more remote clusters, and also has solar and wind energy. This is why Zhangjiakou is attracting significant private data center investment in addition to those of state-operated telecoms, making it a relative success compared to some of the other more far-flung clusters such as in Gansu and Ningxia, or those in Xinjiang (which are not officially part of the EDWC plan).

The EDWC project was conceived to coordinate the growing energy needs of data centers with renewable energy buildout. However, the process of data center and energy permitting has turned into a new arena of political contestation. Some data center operators bribe their way to obtain permits for building from local governments, fueling a speculative boom that may outstrip the capacity of renewable energy in the areas where they are built. Even as data centers are built adjacent to renewable sources, they actually draw heavily on dirty energy like coal or gas-fired power plants, which provide a more reliable energy source required by data centers. According to a recent report from Xinhua , Zhangjiakou’s data centers used 1.859 billion kWh, up 43.47% YoY—-”green electricity” accounted for 574 million kWh, about 30%.1

By some measures the effort has been a success—according to official statistics released by China’s National Data Bureau 45.5 bn Yuan ($6bn) of capital has been invested in data centers via the plan.2 However, other reports suggest that many data centers lie underutilized or even empty. When the effort was conceived, the idea was to “optimize the layout of China’s data centers.” The plan sought to avoid overbuilding of data centers in crowded urban areas where land and energy prices are higher and concentrate data centers where there was ample cheap and renewable power supply. But in fact, the EDWC policy has encouraged many cities beyond the eight official clusters to pursue data center development. Real estate companies and other conglomerates have tried to grab a piece of the data center building frenzy, leading to creation of companies that have little experience in data center buildout. The result is a familiar story of overbuilding and overcapacity—time will tell if data center demand will catch up with the overinvestment, or if the frenzy will end similarly to China’s real estate boom, with empty and/or underutilized facilities.

In 2022, four agencies jointly issued a plan called Eastern Data Western Compute 东数西算, which aimed to boost China’s computing power by constructing new data centers in eight designated clusters across the country.3 China’s western provinces are rich in energy: both traditional oil and natural gas, as well as renewables like wind and solar. Meanwhile, most data centers had been built in the prosperous eastern metropolises, to be nearer to most customers. Latency, a measure of the time it takes for a packet of data to travel between a user and the server and back, is highly correlated with distance: closer computing facilities enable lower latency, and faster computing. But the western hubs would be suitable for certain data operations: cold storage and backup. With the advent of energy-intensive AI model training in the wake of ChatGPT’s release, suddenly inland data hubs had a newfound purpose as ideal sites for intensive model training operations. As many of you know, I’ve been covering the EDWC policy for a while, but this year many English language outlets began to pay attention—writing about newly built data centers in Xinjiang, Qinghai, and Wuhu—which the FT called “China’s Stargate”, alluding to the massive OpenAI backed data center being built in Abilene Texas.

The EDWC plan built on the initial success of Guizhou, which began attracting data centers by major Chinese tech companies around 2013, and later Inner Mongolia, which sought to replicate Guizhou in the North. Both inland regions are blessed with suitable cooler climates, and energy: Guizhou with hydropower, and Inner Mongolia with wind and solar. In 2022, the EDWC plan was released, adding six additional clusters to a national framework. Four of these (Shaoguan, north of Guangzhou; Wuhu West of Shanghai, Zhangjiakou outside Beijing; and Chengdu-Chongqing) are located in or near major metropolitan areas, while four of them were in remote Western regions: Qinyang in Gansu, Hohhot (Inner Mongolia), Guizhou, and Ningxia.

Meanwhile, many cities outside these began building data centers. Datong is a city in northern Shanxi province, just southwest of Zhangjiakou. While no cities in Shanxi 山西 province are part of the official EDWC plan, the province’s climate, location, and cheaper land and energy costs make it an ideal location for data centers like its larger neighbor to the north, Inner Mongolia. Chindata, a leading Beijing-based data center operator, built a massive hyperscale facility in Datong. Other cities in Inner Mongolia have cultivated their own data clusters. Ulanqaab 乌兰察布is a prefecture-level city in Inner Mongolia, located about halfway between the provincial capital Hohhot and Hebei’s Zhangjiakou. It has developed a big data industrial park with large cloud computing facilities, including Apple’s second data center in China. Even within Inner Mongolia, the rise of two cloud computing hubs is indicative of the decentralized rollout of data center development, seen by local governments as a new engine of growth. Specialized data centers in Xinjiang and Tibet have also attracted media attention recently. While these regions have high solar and wind potential, their distance from urban areas makes them relatively unattractive for most types of data centers, but they could be used to train AI models, which does not require lower latency needed for real time computing.4 Desert data centers in remote Yiwu, Xinjiang got attention from Bloomberg earlier this year, but in reality China’s data center boom is mostly still happening in areas closer to major cities—in places like Zhangjiakou 张家口which offers both proximity to Beijing, and benefits of cooler climate and abundant renewable energy. Getting trained engineers and reliable infrastructure to operate data centers in Inner Mongolia and Gansu is hard enough, let alone in remote Xinjiang.

Data centers are promoted as drivers of digital industry and remedies for the “digital divide” in the underdeveloped and poorer areas in which they are being built. However, the construction of data centers offers few of the benefits of productive industries such as factories. There are few jobs and because data centers don’t directly produce anything, they don’t offer much in the way of stable tax revenues for local governments. They don’t lead to many spillover industries, as the main beneficiaries are users located in main urban areas. Local governments have given land subsidies and tax breaks for data center construction, with little to show for it in return.

Securing Energy and Water

Some municipalities can generate revenue by selling credits for renewable energy supply used to power data centers. In the case of Zhangjiakou, I was told by a researcher who had met with local officials there, that they city was essentially selling quotas for wind power (风电指) to generate revenue to pay down the city’s debt, which had accumulated from years of overinvestment following the 2022 Winter Olympics. “The city is essentially mortgaging its future,” he told me. “The city has to use the revenue to pay down its existing debt, and doesn’t get much benefit from the data centers directly.” The city can sell its credits to energy operators, which build the renewable power that is being used to justify the data center boom. The wind power quotes are based on a city’s wind power potential, so Zhangjiakou was allotted a higher amount.

Huailai County怀来县 is ground zero for Zhangjiakou’s data center boom. Adjacent to Beijing’s Yanqing County 延庆县, home to the famous Badaling section of the Great Wall, Huailai has an ideal climate for data centers (and wine grapes, as well). Proximity to Beijing offers low latency for real-time computing needs. The 2022 Winter Olympics, which were mostly hosted in Zhangjiakou, resulted in a boom of investment in infrastructure and housing. Like many cities in China, Zhangjiakou is heavily in debt. New housing flats constructed a few years ago are mostly empty, some of them purchased by Beijing residents looking for a cheap place to retire or visit in the summers, when the area is cooler than the capital. Wind turbines blanket the area, some of them dotting the shores of Guanting Lake, a large reservoir. But a local tells me these projects mostly pre-dated the data centers and provide energy to Beijing. “There hasn’t been much economic impact from the data centers. They employ some local people, but mostly in lower jobs—technician or security guards. Most of the engineers are college graduates from elsewhere, or Beijing. A high-speed rail station near the area is just a 40 min train ride to Qinghe Station, in Beijing’s Haidian District-where the country’s top universities and an important tech industry cluster is located.

Data centers consume huge amounts of electricity. The approval of data centers is supposed to be conditional on the securing of power supply from the local grid. According to one person I spoke with, a family friend in Baoding (a city in Hebei just southwest of Beijing) was trying to bribe their way into securing a permit for a new data center, but was still unable to get approval for the energy permitting. The local branch of the NDRC (发改委) is supposed to approve data center construction in each municipality. In July of this year, Xi gave a speech earlier this year where he cautioned against cities rushing to build data centers.5 So there is official recognition that the data center buildout may have gotten ahead of itself.

The EDWC plan was premised on the securing of renewable energy supply. In practice, however, many of the data center hubs’ energy needs exceed that offered by renewable power. Even in Zhangjiakou, a hub for wind and solar energy, the data center draws only part of its supply from wind and solar. Much of it comes from a coal-burning steam-powered turbine facility in the county of Xuanhua, operated by central SOE Datang Group.6 In Hohhot’s data center cluster south of the city, a new steam-turbine power station was built by a Beijing-based power company near the cluster to secure supply.7 Wind and solar do not yet provide the constant reliable power that data centers require. Until improvements in battery storage or efficiency, the promise of green data centers may be more “greenwashing” than reality, as data centers suck up power that requires stable if dirtier sources as coal and natural gas.

Water is also required to cool data centers. The Zhangjiakou data cluster is in an arid region, although the Huailai cluster is close to the Guanting Reservoir. But according to conversations with villagers in the area, water used to cool data centers has to be sourced from groundwater, not the reservoir. This means the cooling demands of data centers could exacerbate groundwater depletion, which is already a significant issue in North China.8

Different Types of Data Center Operators

Unsurprisingly, much of the investment in the EDWC computing hubs has been from the three major state-owned Telecoms (China Telecom, China Mobile, and China Unicom)—see my figure from data compiled in 2024. But there are two other main categories of data center operators building data centers: some are built by i.) leading cloud companies (Ali Cloud, Tencent, and Baidu), and many by ii.) third- party data center operators that lease their space to smaller companies (called the “co-location” model)—such as Chindata 秦淮数据, 21 Vianet 世纪互联, and Shanghai-based GDS 万国数据. ZData (中联数据) is another growing third party data center operator with a new center in Hualai County as well as in Shanxi province.

These companies are now expanding abroad and have to carefully balance maintaining their China businesses and good relations with the government while expanding abroad—which often requires them to hide their connections to China state entities or create new entities. For example, Chindata, whose largest shareholder is Bain Capital, restructured itself. A holding company based in the Cayman Islands since 2018 owns some of its overseas investments such as Bridge Data Centers, which has been building in Malaysia.

Their Beijing-based subsidiary oversees its China projects, most of which are in Northern China. GDS, which is partially backed by Shanghai’s Zhanjiang high Tech Park, but also owned through a Cayman Islands holding company, has pursued a similar strategy as it expands overseas with new data centers in Malaysia and Singapore. GDS has also evolved from being a mere data center operator to an advanced AI infrastructure provider.

The EDWC boom has drawn in a new cohort of data-center operators beyond the telecom SOEs and established cloud platforms. These are often started as by real estate developers or other conglomerates that have little experience building data centers but are trying to get a piece of the cake. For example, one of the largest data center facilities in Zhangjiakou (a planned 1000 Mw “data industrial park”) is being built by a company called Hoyinn 合盈 The company is majority owned as a subsidiary of the Tianjin-based Zhengxin Group 正信集团, which has primarily been a residential real estate developer.9 Beijing-based Haoyang Group, began in 2021 as a relatively small player with a data center in Langfang, located between Beijing and Tianjin, and is now building a hyperscale facility in Thailand.10 The Chairman of Haoyang, Lai Ningning (赖宁宁), founded CNIX, an internet exchange company in 2016. The company is partly backed by U.S-based Apollo Capital.

Conclusions

Zhangjiakou’s data center cluster is already one of the more developed among the eight—and because of its proximity to Beijing and favorable energy potential, offers an ideal mix of conditions for data centers. But even in Zhangjiakou, data centers have, like elsewhere, failed to deliver significant economic growth. The city remains highly indebted due to years of excess infrastructure buildout both leading up to and following the 2022 Winter Olympics. The dynamics of energy policy and corporate investment in Zhangjiakou are themes that can be seen in China’s broader data center buildout, which has already led many observers and even Xi himself to caution against the rush towards data centers by local governments. But for now, there seems to be no sign of a slowdown…

I landed in Beijing’s capital airport on a cold evening in October. Inside T3, which was completed just in time for the 2008 Beijing Olympics, bright towering illuminated advertisements from Huawei, Ali Cloud, Bytedance, and other leading tech firms greeted passengers as we got off the plane—totems to the age of AI. China and the U.S. are leading players in this global race (some would say bubble) toward “artificial general intelligence,” or at least what we perceive to be artificial intelligence. This trip in China was characterized by visits of some of China’s leading technology coupled alongside a persistent pessimism toward the overall economy and livelihoods for everyday citizens.

In Shanghai, I visited an exhibition of consumer-driven AI technologies at the West Bund, a new district south of the city center home to leading tech companies including Alibaba 阿里巴巴, Sensetime 商汤科技, and Tencent 腾讯, as well as new museums like West Bund and Long Museum. Inside the exhibition hall, I was greeted by a smattering of robots, some of them cute and furry, others more utilitarian. A robotic coffee machine processed a picture of my face and printed it on the top of a latte, all for 20 Yuan ($2.80). A robotic ping-pong arm rallied with a girl as she returned each serve. Another small tabletop robot produced by Sensetime played the Chinese game of go with an older woman locked in concentration. Nearby, a humanoid robot from Unitree 宇树 danced and gyrated. Meanwhile, a man surreptitiously operated a soccer-playing robot as the crowd watched from the other side, assuming it was kicking a soccer ball of its own volition. The sheer variety of products and technologies on display are evidence of the enormous creativity and experimentation going on in China’s tech space, and one where “AI” features and physical products are increasingly intertwined.

“Since the advent of ChatGPT, there have been growing reports of people using AI chatbots for psychological counselling or cures for loneliness. In China, this has taken a cuter and more physical form”

Another category of products on display were what I would describe as “cute companion tech”, dolls or plushies equipped with intelligent conversational abilities to offer comforting words of solace for overworked niuma 牛马1 after they return home from grueling hours in China’s 996 culture. A capybara plushie was advertised as an “AI companion,” although the device wasn’t working so I couldn’t get to test it, even as I toyed with the idea of buying it for my nephew. Many of the products on display promised some form of emotional companionship or mental health diagnosis. A pendant necklace equipped with sensors for body heat and other diagnostics was promising the world’s first emotional sensing AI necklace. Since the advent of ChatGPT, there have been growing reports of people using AI chatbots for psychological counselling or cures for loneliness. In China, this has taken a cuter and more physical form, with AI-powered companion devices or plushie animals equipped with AI-powered conversation ability, promising to cure loneliness or other mental ailments. An attendant showed me a humanoid robot that had soft muscle-like material that was more tactile and gave off heat, to better replicate human qualities compared to the rather spindly and somewhat creepy Unitree robot. The man at the display told me they were intended as companions for elderly or disabled.

Just as with Silicon Valley techno evangelists, the idea that more technology is the answer to the global loneliness epidemic, which has itself been fueled by social media and solipsistic tendencies of phone addicted youth, is alive and well in China’s AI age. Addressing the loneliness epidemic of modern life with more tech is like fighting poison with poison (以毒攻毒). Meanwhile, neither societies have good answers for the really hard questions—what should unemployed graduates should do with their lives, and deeper questions like what kind of life is worth living when work is no longer the main purpose of our existence? On the other hand, many of the products I saw are market responses to the reality of China’s ageing population. In a country long defined by filial piety, how do children make sure their elderly parents living far away are taken care of?

Many of the products on display were gimmicky, to say the least. A small humanoid robot kicking a soccer ball around a small pitch was being manipulated by an attendant, his fingers manipulating a hidden controller beneath the table to give the illusion of autonomous control. Another “cosmic AI companion” which was due to be released in a few months, would offer astrological advice and fortune telling. The cheery young attendant who approached me in clear English had just graduated from UC Berkeley and had returned to China to work in this small AI startup.

A smaller section of the exhibition floor in the rear featured industrial applications: an AI data platform that would optimize energy use for renewable power in large-scale industrial parks or factories was evidence of the integration of AI with China’s growing capacity in renewable energy and infrastructure. Nearby a woman dressed in an ethnic minority costume was standing by a booth from a small NGO advertising themselves in “rural education power by AI” which she told me could develop specialized teaching modules based on local needs in remote villages that lack adequate educational facilities. The idea was interesting but I didn’t quite understand how such a solution would be scaled or integrated into China’s vast rural areas, and the woman was scant on details of actual successful case studies or uses.

Despite some of the questionable products on display, the vast array of products was evidence that China’s AI sector is seeing a flurry of experimentation in AI-powered consumer tech. Whether or not these products are true examples of “AI” is another question. A friend of mine who is finishing a PhD in economics at Tsinghua tells me 绿色漂洗 luse piaoxi “green washing” is a popular topic in Chinese academic discourse recently, and surely we can add 数字漂洗 “digital washing” and 人工智能漂洗 “AI washing” to the mix. Companies promote themselves as being in AI to get funding and government subsidies. But China’s manufacturing prowess and its ability to quickly prototype physical products is a distinct advantage in its AI innovation ecosystem compared to the U.S. where LLM-powered chatbots have mostly been adopted by enterprises business services and for personal productivity.

As in the U.S, a lot of the money and innovation in China’s AI space is still flowing into a handful of leading technology companies. A friend who recently left a small startup to work on integrating AI solutions into Alibaba’s urban mobility applications, told me that they were trying to develop AI-powered transportation recommendations based on aggregating data from the hundreds of Chinese cities where passengers use Alipay to board subways or buses. Ali has access to a staggering amount of data on travel patterns in most large Chinese cities which, if utilized properly, could deliver customized travel suggestions or more efficient routing recommendations. Like the ride hailing app Didi or its global cousins Uber and Grab, Ali is poised to become an urban mobility company just because of its access to this data alone. But these solutions haven’t begun to generate profit for Ali yet, and the company is still investing in research and product development in order to find ways to monetize this data, using AI models as an accelerant.

Even as the AI race ignites a new wave of speculative investment, there is a sense that many other sectors are still struggling. A former colleague who works in Shanghai for one of the country’s leading real estate developers spoke of the still-dormant property market. The company she works for is trying to pivot to digital services in its existing properties since they are building fewer greenfield communities. While she had a decent position advising the company on digital solutions, she herself considers moving to one of China’s tech giants since they are the only ones expanding and seeing real growth these days. “A few years ago, a lot of people like me with overseas degrees in Shanghai would have worked for MNCs because they paid well, but many of them have left China.” Another Chinese friend who had studied in the U.S. and was working for a foreign environmental NGO in Beijing that is helping Chinese companies decarbonize was considering moving abroad, perhaps to Singapore, and said there were few similar non-Western organizations in China that paid a comparable salary for the kind of work he was doing.

Like in the U.S., the AI race/bubble is sucking investment and energy (both in kilowatts and zheng nengliang) out of other sectors and concentrating wealth and capital in a few leading firms. This is a commonality that is often lost in discussions of U.S. and China tech competition—AI promises wider benefits but seems to be concentrating wealth and resources even further, accelerating inequality and resentment in society that were already bubbling, both during China’s “Gilded age” boom years, and in the populist backlash toward the elites that has transformed American politics.

China’s government is aware of these risks and has pushed policies to make AI a tool for industrial development rather than a thing unto itself. There is a focus on application of AI to industry, production, and the 实体经济 shiti jingji “Real economy.” Under Xi, the government has taken a skeptical attitude to consumer internet and favored investment in “hard” industries like manufacturing, new materials, infrastructure. But this rhetoric is juxtaposed with market dynamics in which China’s leading tech players, with their existing advantages in data and software engineering talent, could be the ones to benefit most from AI, applying large language models to make their existing operations or services more efficient.

Despite this, there are other efforts to use AI and technology to address China’s ageing population and deeper social issues. In the Lujiazui area of Shanghai, not far from the glittering skyline that marked the city’s arrival as a global city, I visited a Community Center built by the district government. Shanghai has pioneered these community facilities, which often include a low-cost canteen with discounts available for elderly residents, party cadres, and veterans. Some of this policy began under the city’s previous Party Secretary Li Qiang, who was elevated to China’s #2 Premier following his faithful implementation of the Shanghai lockdown, which is a traumatic memory for many Shanghainese and those who lived through it. This center also had numerous facilities for intelligent healthcare and elderly services. There are various “smart products” on display: a smart couch that can move to help elderly get up, a smart bed. There are rooms one can book here to stay, and receive services like health checkups, treatments, and other activities.

There are services offered by a company called Guorong Leyang —a joint venture of Shanghai’s Fosun Pharmaceutical and Sinopharm, the state-owned pharmaceutical giant, that is providing many of the services on offer. For a few hundred yuan an hour, you can hire a young assistant to accompany your grandmother or grandfather to the hospital to help her navigate her treatment. It’s part of the growing “silver economy” in China—turning the country’s sharply ageing population into yet another area of economic growth. The community center is a showroom for digital products and services of the company. One of them is an automated health detection screen, which primarily focuses on using traditional Chinese medicine diagnostics. It scans my tongue, which is seen in Chinese medicine as a window into the body’s overall health, for a report, and then delivers it to my phone via a QR code. I have too much moisture in my body, according to the report. I need to eat more mushrooms, and avoid keeping my windows open at night. I visit an actual Chinese medicine practicioner to try aijiu and baguan a few days later, and he tells me something similar. So I guess the screen wasn’t so far off, at least as far as translating the principles of Chinese medicine zhongyi into a digital form.

Despite all of the technological progress and Party-led aspirations for technological self sufficiency and even supremacy, there is still the inescapable anxiety and pessimism pervading conversations with people. Innovations in clean energy, robotics, and electric vehicles have improved life in many ways, and certainly aided China’s transition to a cleaner development model. But everyday workers and citizens have not seen much improvement in their own fortunes during the last few years. The property glut continues to drag on the overall economy, as Chinese urban citizens had relied on property assets as a major form of investment during the boom years, lacking other options for investment either in China or overseas.

In a post last summer I wrote about “neijuan”, 内卷 the term that has broad meanings but has come to be a shorthand for the over-competition and overwork that characterizes contemporary life in China. Commonly translated as “overcompetition” when describing the economic situation, the term also has spiritual and cultural connotations, a general “turning inward.” While the China of the early 2000s saw a booming interest in the world, Xi’s China is unmistakably turning more towards its own traditions and itself. A friend working in Shanghai told me some of her younger post 90s generation cousins (roughly equivalent to Gen Z in the U.S.) have little interest in going abroad or in travelling, which was much different from her experience just 10 years before.

Turning inward can also be felt in the spaces urban residents seek out after work. I visited an onsen or a 汤泉 tangquan in the Sihui area of Beijing just East of the CBD— a bathhouse where people spend hours on weekends or after work. China has long had a culture of public bathing and onsens, particularly in the colder North. But in recent years, these facilities have become less about bathing and more about eating, conspicuous lounging, and the influencer or 网红 wanghong economy. For owners of the bathhouses, there is not much use in spending money on designing the bathing area, where phones are not allowed, and therefore not promoted on xiaohongshu—China’s version of Instagram that has quickly come to dominate social and urban life. After bathing, you retire to the upstairs eating areas, offering a cornucopia of tropical fruits, hairy crabs, ice cream, popcorn, and other delicacies.

The design here is exuberant and aimed at influencers, the more to go viral on social media. The influencer economy has transformed how people use urban space. On the bund in Shanghai, I watched a rural migrant breakdancing by himself, livestreaming away probably on Kuaishou or Douyin—perhaps promoting a product or just communicating with his followers on social media. During a thirty-minute subway ride in Shanghai I sat next to a young man in a black jacket as he moved his thumbs furiously, playing a game on his cellphone, not once looking up from his unbroken concentration at any of the other passengers around him.

While these forms of digital anomie are not unique to China, they do seem to be particularly evident amidst the hypercompetitive and stressful rhythms of everyday life in urban China. While American discourse on China is dominated by fear of China’s growing technological and military rivalry, the reality of daily life in China is a bit less glamorous and more mundane. Technological convenience is indeed a part of life in China, ranging from ubiquitous mobile payment systems to robots and electric vehicles. But there is also a sense that China may be following a path of “Japanification,” a society that turns inward even as it achieves remarkable technological progress and advancements in living quality and convenience. More than anything, this trip was a reminder that despite the tendency to view China through lens of exceptionalism or hyperbole, the problems facing China’s people are in many ways shared with those in the U.S. and elsewhere. The dominance of tech giants over urban life, a speculative race towards artificial intelligence and a shrinking pool of good paying entry-level jobs for college graduates, and a turning inward—can be felt in many places. China will surely chart its own course as it deals with these problems, but its always good to do a reality check on the hyperbolic and narrow-minded focus on tech rivalry that pervades U.S. coverage of China these days.

Sometimes CAS is imagined as consisting primarily of its sprawling main campus in the Haidian District of Northern Beijing. Zhongguancun Hi-tech park emerged as one of China’s key IT innovation districts due to support from nearby CAS. I have done a quick analysis, based primarily on peering into publicly available investments structure of CAS, to reveal the depth of CAS’s sprawling nationwide network of research institutes, companies, and subsidiary investments. It counts 67,000 staff and 50,000 researchers. In the last ten years, CAS has become the world’s preeminent research institute according to Nature’s index of research output, measured in terms of highly cited publications and researchers, surpassing the UC system and Oxford/Cambridge.2 But these statistics do not account for the subsidiary investments, which count among them some of China’s leading-edge companies in AI, optoelectronics, and semiconductor development.

The research of CAS has always been closely tied to China’s state objectives, including close cooperation with defense technology and military-civil fusion. But in the last 10 years or so, CAS has evolved further a traditional national research organization into a diversified state-owned conglomerate, with holding companies, venture capital funds, and numerous subsidiaries. The CAS operates labs in cities across China, which increasingly function as nodes within regional and urban R&D ecosystems. Several institutional restructurings have transformed CAS into something of hybrid national research institute cum diversified state-owned enterprise. According to a 2024 report by Georgetown’s CSET, CAS has played a role in founding notable companies like chip-maker Cambricon, supercomputing firm Sugon, iFlytek (an AI firm), and Lenovo.3. The CAS is also involved in the development of several “national-level Science Parks”, 国家综合性科学城 large-scale new R&D-focused economic zones in Beijing, Shanghai, Shenzhen/Dongguan, Hefei, and Chengdu, anchored by CAS affiliated institutes.4

National Innovation Ecosystems

The relationship between nationally-funded scientific research and commercially-oriented innovation has been a key debate in many countries at different periods of modern history. In economic geography and innovation, the concept of “national innovation system” (NIS) has been a dominant theme. Lundvall (1992) is credited with coining the term, noting its intellectual roots in the 19th-century ideas of economist Friedrich List.5 In general, an NIS refers to the network of institutions and relationships in a nation that drive technological innovation. Freeman (1987) defined an NIS as “the network of institutions in the public and private sectors whose activities and interactions initiate, import, modify and diffuse new technologies”6

In the U.S., this debate crystallized during World War II, with efforts to boost the allied war effort with defense-related research in radar, nuclear, and eventually what would become semiconductor technologies. During this period, institutions such as the NSF (founded 1950), the National Institutes of Health, the Department of Defense’s research offices (e.g. DARPA, founded 1958), and NASA (1958) were established or expanded, alongside a network of national laboratories and research universities . This public infrastructure — inspired by Bush’s linear model linking basic science to applied outcomes — fostered major innovations (from semiconductors to the internet) while largely leaving product development and commercialization to the private sector .

Commissioned by President Franklin D. Roosevelt in 1944 and submitted to President Harry Truman in 1945, MIT Professor of Electric Engineering Vannevar Bush’s Endless Frontier report called for basic research as a public good, and envisioned partnerships between government, universities, and industry — a so-called “triple helix” model of R&D. The U.S. NIS evolved with substantial state support for knowledge creation, but a comparatively decentralized, market-driven application of that knowledge. Bush’s ideas formed the basis of a bipartisan consensus on the importance of federally funded research that is now under direct assault by the Trump Administration’s attack on federal funding for university research.

Japan and China offer models of state-driven innovation ecosystems in an East Asian context. Freeman showed that Japan’s economic catch-up was bolstered by state coordination of technology policy . The Ministry of International Trade and Industry (MITI) in the post-1950s era acted as an innovation orchestrator – guiding industrial priorities, fostering close government–business cooperation (especially with large keiretsu conglomerates), steering financial and educational resources toward key sectors. The Japanese NIS included a strong central push for modernization, heavy investment in technical education and training, proactive importation and improvement of foreign technologies, and public-private collaboration to diffuse knowledge throughout industry .

Efforts to create a national innovation system in China have been underway for many decades. Some highlights of this include: the “863 Program” (National High-Tech R&D Program, 1986). Launched in March 1986 after a proposal by prominent Chinese scientists to Deng Xiaoping, 863 focused on frontier technologies like IT, space, biotechnology, new materials, and automation. The Torch Program (1988) aimed at commercializing technology and fostering high-tech zones (e.g., Zhongguancun in Beijing). It supported technology transfer from research institutes to enterprises, promoting science parks and incubators. In 1997, the “973 Program” (National Basic Research Program) sought to strengthen basic science capabilities and link fundamental research to strategic sectors. It complemented the applied focus of 863 by funding long-term, large-scale basic research projects.

CAS as State-led Innovation Platform in the New Era

Developing China’s basic R&D and innovation ecosystem has been a major focus of Xi’s tenure. Made in China 2025 (2015) targeted advanced manufacturing and reducing dependence on foreign suppliers. The current “Whole-of-Nation” strategy in the 14th Five-Year Plan (2021–2025), stressed self-reliance and support for choke-point technologies. Under S&T Reform Plans (2021–2023), CAS is undergoing restructuring—streamlining labs, funding, and evaluation systems—to further solidify its role as a hybrid innovation platform, embodying both national research ideals and enterprise-driven commercialization. CAS has also been setting up entities for venture capital investment and asset management.“The Chinese Academy of Sciences Holdings ("CASH"; 国科控股) is an entity set up to manage CAS’s commercial and financial enterprises the companies that are invested in by the Chinese Academy of Sciences (CAS). “CASH was approved by the State Council as the first asset management company engaged in the assets operation of a state-level public institution. Since 2008, CASH has participated in Venture Capital and Private Capital (VC/PE) investments as an institutional investor, and actively conducted direct investments centering on China's strategic emerging industries. CASH has over 30 subsidiaries, covering fields of information technology, high-end equipment manufacturing, environmental protection and new materials, investment, and modern services. Well-known investments include Legend Holdings, China Sciences Group, and China Science Publishing Group.” The Chinese Academy of Sciences Innovation Investment Co (Guoke) 国科科技创新投资有限责任公司 was founded in 2015 under CASH to be a vehicle for translating CAS’s scientific research into commercial ventures. It invests in high-tech innovation companies, particularly at early and growth stages—covering sectors like biotechnology, artificial intelligence, IT, smart manufacturing, and life sciences. In 2017, Chinese Academy Of Sciences Venture Capital Management Company was founded.

Investments into new companies has not come solely from the specially created investment and holding platforms at the top of CAS. Specific CAS institutes also have their own investment and holding firms, which have incubated notable companies. Lenovo 联想 was founded in 1984 by Liu Chuanzhi and 10 other researchers with funding from the Computing Institute of the Chinese Academy of Sciences. Lenovo was originally Legend Computer, until the English name was changed in 2004. Legend Holdings Corporation (联想控股股份) remains as a CAS-controlled holding company with diversified investments into numerous companies. CAS retains a 30% stake in Legend (via CASH), and therefore retains indirect ownership in Lenovo itself.7

The CAS lab Changchun Institute of Optics, Fine Mechanics, and Physics helped invest in a new semiconductor lithography equipment manufacturer Gopptix, headquartered in Beijing’s Yizhuang area. Another example of a CAS-held venture capital platform is the the Xi’an/Beijing-based CASSTAR Technology Co., Ltd. (中科创星科技投资), which is partially owned by the Xi'an Institute of Optics and Precision Mechanics, a CAS lab in Xi’an. The company is a venture capital platform that invests in numerous companies, primarily in Xi’an and Beijing. 8 The Xi’an CAS operates with joint investment from Xi’an LGFVs and district investment platforms in the city’s High-Tech Area (Gaoxin Qu).

Other firms with CAS-connections include iFlytek, an AI company based in Hefei that is partly owned by the University of Science and Technology’s innovation holding co. iFlytek (科大讯飞) is one of China’s leading artificial intelligence (AI) companies, known for its speech recognition, natural language processing, and AI education technologies. Cambricon 中科寒武纪科技股份 is a chipmaker focused on AI chips. Founding brothers Chen Tianshi and Chen Yunji studied at USTC, then joined Chinese Academy of Sciences (CAS). CAS funded the startup Cambricon with a seed grant worth several million dollars.9 In fewer than five years, Cambricon has gone from a startup to a company that supplies Huawei and is now listed on the Shanghai Stock Exchange and has been called China’s Nvidia.10

CAS as Regional Innovation Anchor

Besides its main headquarters in Zhongguancun area of Beijing, CAS is actually a sprawling nationwide system with major branches in Shanghai, Guangzhou, and clusters of labs in Chengdu, Xi’an, Ningbo, and even far-flung Urumqi, Lanzhou, Kunming, and a botanic garden in Xishuangbanna. CAS-affiliated labs in specific cities are anchors of growing regional innovation ecosystems. This is particularly evident in the development of “national comprehensive science cities” 国家综合性科学城, a plan put forward in 2016 to build out large specialized campuses for basic research across China—beginning with campuses in Beijing (Huairou), Shanghai, Hefei, Greater Bay Area (Shenzhen/Dongguan), and Chengdu. Many of these are anchored by key CAS institutes. The CAS has also been expanding its academic footprint, with new colleges of CAS University in Chengdu, Nanjing, Xi’an (see below).

The map below shows the geographic extent of major CAS labs and affiliated institutes.

Further research

In later posts, I’ll delve into more detail about some of the research and science parks under development in China and their urban and regional formations.

The Chinese city in the planned economy

One assumption of those interested in Chinese cities is that between the end of the Chinese Civil War and the the Reform and Opening up period (from 1949 to about 1979), Chinese cities were essentially stagnant. But this is not really accurate. In the 1950s as part of the first Five Year plans, state-led heavy industrialization of cities was promoted, influenced strongly by Soviet advisors. Under some of Mao’s more extreme policies like the Great Leap Forward and then Cultural Revolution (1966-1976), Chinese cities stagnated in the face of policies designed to move production to rural areas, or send educated youth to the countryside. Professional city planning virtually ceased and universities shuttered. But this was also regionally dependent—during this period, some areas of the “third front” 三线 like Chengdu, Guizhou, and Xi’an saw increased investment in new factories.

Another misconception is that the legacy of the planned economy has been entirely jettisoned during the reform period. However, this too is not entirely accurate. While many factories shut down during SOE reforms of the 1990s, significant entities have survived and persisted to the present. For example, Chengdu Aircraft Factory is now part of AVIC Chengdu Aircraft Industrial Group 成飞, one of China’s major defense contractors that produces some of its most advanced fighter jets.1 Shougang Group 首钢集团 is the successor to Beijing’s Capital Iron and Steel Factory, and has evolved into a diversified holding and investment company owned by the Beijing government.2 The “fixed” nature of urban land use has meant that much of the land allocated to SOEs during the planned economy period persists in some form 60 years later. Many of the entities involved diversified, leveraging their land holdings to go into property development. Many old factories did shut down, and some of them have been preserved as revived cultural districts, like 798 in Beijing or Chengdu’s Eastern Suburb Memory 东郊记忆.

When I was working on a previous post on Chengdu, I happened upon a map of Chengdu from 1989 produced by the CIA. I was struck by the level of detail. The map provided a clear snapshot of major factories, educational institutions, and government agencies in the city. From these maps one can get a pretty good picture of a particular cities’ industrial mix, its key factories and sectors. I then discovered similar maps of other Chinese cities. The maps show the results of urban industrial expansion in the post 1949 period. The form of the post 1949 Chinese city followed general patterns. The pre-1949 historic cores of Chinese cities dating to Qing periods were maintained but adapted—previous Qing-era government sites were often adapted for government buildings or educational institutes. Some factories were located in the inner cities. But many cities expanded, with railways and new factories forming industrial areas ringing the old historic cities.

The maps also reveal divergence between cities. Historic coastal cities like Guangzhou and Shanghai are little changed from their pre-1949 extents but cities like Chengdu, Zhengzhou, and Beijing saw significant expansion through new industrial areas on their peripheries. Indeed, coastal cities like Shanghai and Guangzhou were generally less favored in the early decades of PRC rule, as they were seen to be former strongholds of the capitalist class that had supported the KMT. Foreign trade slowed significantly, rendering previously open ports less relevant. The maps thus reveal spatial inequalities between Chinese cities—inland hubs were favored for new industrial production under the First Five Year Plan (1953-1957). Of course, the fates of these cities would again diverge in another direction in the 1980s, as investment flowed into coastal economic zones.

Why were these maps produced? The CIA maintains accurate maps of many foreign cities. During the 1980s, the U.S. and China undertook rapprochement, with trade and investment flowing after decades of closure. Nevertheless, the CIA still maintained accurate maps of Chinese cities, particularly defense-related industries. Most likely, these maps helped develop lists of useful targets in the event of war, and may have assisted with other intelligence gathering efforts. The maps generally do not highlight residential areas, but these are assumed to be the grey urban areas in between everything else. This may have overlooked the fact that some areas of factory sites were indeed devoted to housing, as part of a system of “iron rice bowl” or provision of all social and living services by the employer, sometimes also referred to as the “Danwei” or work unit system.3

Beijing

Mao favored an industrialization of the Chinese capital. "Looking out from Tian'anmen, I should see chimneys everywhere,” he is alleged to have said.4 A core idea of Communist ideology was that in capitalist systems, cities are mostly “parasitic” on the countryside—they drain resources from farmers and don’t produce anything useful. Under Mao, Beijing was to be transformed to an industrial city. In the 1960s, the old walls of Beijing were gradually torn down, a symbol of feudalism. However, these maps of Beijing reveal that for the most part, the inner core of Beijing (within what is now the second ring) was not the location of new industrial production. Large industrial complexes were built to the West—the Capital Iron and Steel Corporation 首钢 in present-day Shijingshan, No 2 Chemical Works in the southeast, and a factory complex in the northeast comprising Wire Communication Equipment Plant 738, Electron Tube Plant 774, radio equipment plants, and the 798 military plant, which has been redeveloped into an arts and cultural district. Meanwhile, the inner core of Beijing was maintained as a residential area, with schools, government buildings, and hospitals dotted throughout.

Shanghai

Shanghai had been China’s most prosperous and cosmopolitan city from the end of the Qing into the 1930s, the “Pearl of the Orient.” Under the PRC, the city stagnated. Long a hub for foreign trade and capitalist production, Shanghai was not favored in early PRC industrialization plans. The maps of Shanghai reveal this. The urban extent (dark grey) of Shanghai as shown in the CIA’s 1982 map is roughly what it was in 1949. A map of 1927 when the city was still administered as a foreign treaty port, split up into concessions, is roughly what the urban area is in 1982. The area to the north of the city has expanded somewhat, with a few large-scale factories, with the exception of Steel Plant No 3 on the south bank of the Huangpu River (the site for the 2010 expo), and the Shanghai Iron and Steel Plants 1 and 5, located in the north of the city. Small textile factories are scattered around Suzhou creek in the city’s northwest, and in the Hongkou District, which had been under Japanese control.

Guangzhou, 1984

A 1984 map of Guangzhou suggests that like Shanghai, this southern coastal trading hub was similar in urban extent from around 1949. However, there are numerous universities found in the northeast—what is mislabeled as Hunan Polytechnic should be South China Institute of Technology (华南工学院) now South China University of Technology 华南理工大学.5 Zhongshan University 中山大学 is just south of the river.

Chengdu, 1989

The southwestern city of Chengdu was a much smaller version compared to its current sprawling extent. But already in 1989, a large industrial zone stretched to the Northeast of the old city. The old imperial-era city, long the administrative hub of Sichuan, was mostly residential but with numerous colleges and government buildings. According to Julie Miao, the most important factories located in the Northeast of the city due to favorable winds, and higher elevation.⁶ Factories such as the state-run Hongguang Electronic Tube Factory 成都国营红光电子管厂 that made cathode ray tubes for TVs and other electronic devices were some of the city’s important industrial anchors. Chengdu Aircraft Factory (now AVIC Chengdu Aircraft Industrial Group / 成飞) was founded in 1958 (the location in Shuangliu District is SW of the below map’s extent). Today, one of these factories was turned into a new cultural space: Eastern Suburb Memory 东郊记忆—Chengdu’s version of Beijing’s 798. During the Third Front (三线建设) campaign (1964–1980) heavy defense-critical industries were relocated to China’s interior under the assumption they would be secured in the event of an invasion by either the Soviet Union or the U.S.6 In this process, aviation, electronics, machinery, and heavy equipment plants backed by Beijing ministries were brought to Chengdu.

Wuhan, 1974

Wuhan, comprised of three cities located in central China, is an important inland port and industrial center. China’s first North-South railway line crossed the Yangtze River here, making the city a critical node in China’s railway network. The city’s largest factory was the Wuhan Iron and Steel Plant, inaugurated in 1958, employed tens of thousands of workers. In the 1980s and 1990s, the workforce was typically around 80,000 to 100,000 employees. It’s visible in the inset of the map below, located northeast of the city center on the Wuchang side. The factory is still running today, under the control of the Wugang Group 武钢集团有限公司, but in 2016 merged with Shanghai-based Baosteel to form Baowu Group宝武集团, the world’s largest steel producer.

Summary: Local state as venture capitalist: the Rise of the “Hefei Model”

In 2015, China's Premier Li Keqiang visited his hometown of Hefei in Anhui to promote an ambitious vision for China to dominate industries of the future. The plan Made in China 2025 sparked anxiety in countries around the world, particularly in the U.S. where it was cited by officials in the Trump administration as justification for newly aggressive policies that have lead to U.S.-China decoupling. Anhui has long been a poor backwater in China—the term 安徽保姆 Anhui Baomu connotes the fact that many domestic helpers in China hailed from the largely rural province. In recent years the city of Hefei has undone this stereotype. And it has done so by implementing the Made in China 2025 vision for excelling in manufacturing in strategic sectors: semiconductors, EVs, and optoelectronics. In 20 years, Hefei’s GDP ranking shot from 80th to 20th in the country, its fiscal revenue increased 33 times, and its population increased from 4.38 million to more than 9.3 million, and its economic output has doubled to $140 bn1. How did the “Hefei model” achieve this?

In 2024, the New York Times reported on the city, calling the Hefei model as “Using government money to buy newly issued shares in manufacturers and start-ups that need cash.”2 Xinhua (state owned media) defined the model in a more process-oriented way: “State capital is directly invested in leading enterprises of emerging industries that are viewed as strategically important but often considered by private investors as too risky or costly. The state capital gradually exits from these enterprises after their market prospects become clearer and they attract interest from non-government investors…State investment returns are then used to support more projects in the industry to cultivate a resilient industrial chain. A healthy industrial ecology is developed in this process, drawing more outside investment, technologies and talent to form a virtuous cycle.”3 But the Hefei model is a bit more complicated than these explanations suggests. Let’s take a deeper dive at the specific entities and their investment strategies. First (1.) I’ll explore the ecosystem of government entities and investment funds comprising the “Hefei model” followed by (2.) several cases in particular industries the city has concentrated on: LCD screens, electric vehicles, and DRAM memory chips. Within each case I look at: a.) institutional arrangements, as well as b.) spatial clustering of firms in industrial zones.

A state-led venture Ecosystem

There are three main state-owned investment and financing platforms : Hefei Construction Investment 合肥市建投集团, founded in 2006 which has the largest capital among the three and focuses on industrial infrastructure (as most LGFVs) as well as electric vehicles, tourism, & transportation. The Hefei Industrial Investment (合肥市产业投资控股) was started in 2015 focuses on emerging sectors like memory chips, while Hefei Xingtai 合肥兴泰 functions as a local financial holding platform and dates to 2002.4 There are around 12,000 of these LGFV entities in cities and other local governments around China. They can come in different flavors, but often feature something like “urban investment/construction company” 城市投资建设公司 in their name.5 So LGFV are by no means unique to Hefei, but they have primarily been associated with land and infrastructure development. Hefei’s main LGFV (Hefei Construction Investment Holding Group 合肥建投集团 or HFJT founded in 2006, is among the largest in China with total assets of 613.584 billion yuan. ($83 bn USD). Like other LGFVs, Hefei’s has also invested in infrastructure and construction projects such as new economic zones.6

What is unique about Hefei’s is that its LGFVs have moved beyond the traditional projects (like infrastructure) to take on the role of venture capitalist, investing in both mature and early-stage firms in strategic sectors. The differentiating aspect of Hefei’s approach seems to be aggressive efforts to woo particular companies and take direct ownership stakes, which usually is premised on that company making a significant investment in Hefei, followed by the government selling its stake at a later time (sometimes quite fast, in 3 years or so)— the company’s production in Hefei remains, and the government has expanded capital to re-invest in future projects.” The investments of Hefei’s funds are extremely targeted, often for the express purpose of helping companies set up specific manufacturing lines for specific products. The state-led venture capital ecosystem in Hefei comprises the entities listed above, as well as provincial platforms. Often, financing vehicles from specific economic zones where the company will set up are also part of the capital mix.

Another key element of the Hefei model is in-house government expertise. According to The Economist, Hefei’s government “created groups of firms in 12 industries, including semiconductors, EVs, quantum sciences and biotechnology. This “chain leader” system 链长制 empowers a government official who oversees big-picture planning for each industry.”7 The local government has maintained in-house expertise which enhances its ability to make good decisions about which industries to invest. There is also close contact between the city and the University of Science and Technology (中国科学技术大学) one of China’s top universities directly under the Chinese Academy of Sciences, which was relocated to Hefei in 1970 during the Cultural Revolution. The university is particularly strong in chemistry and nano-materials, with significance for many manufacturing sectors. “A cadre from the municipal agency has served as a project excavation manager at the Advanced Research Institute of USTC for 7 years. His main task is to help scientific researchers sort out patent results and connect with market demand.”8 The USTC has helped Hefei rank as one of the world’s top 20 cities in terms of scientific research output, according to an index produced by Nature in 2024.9

Guiding Funds

Many of Hefei’s guidance funds are administered under management of LGFV such as HFJT Hefei Jiantou 合肥建投 and Hefei Industrial Investment Hefei Chantou 合肥产投 to focus on specific industries.10 Some notable ones include: The 合肥芯屏产业投资基金 Hefei Xinping Industrial Investment Fund or the “Chip Screen Fund” set up in 2016 to invest in strategic industries, focusing on semiconductors and screen displays. The 合肥瀚和投资合伙企业 Hefei Hanhe Investment Partnership is a joint operation between the HFJTJT and Anhui Province. Other funds include Jianxiang Fund 建翔基金, and Xincheng Fund 鑫城基金 (of the Hi-Tech Zone Admin Committee).

Case 1: LCD Screens and BOE Group

In the early 2000s, Hefei had already established itself as a home appliance manufacturing center, with factories of Midea, Hitachi, and Gree (格力). However, the reliance on Japanese and Korean companies for screens was preventing the city from moving up the value chain. In 2008, Hefei's government spent US$3.5 billion for a controlling stake in BOE, a Beijing-based LCD screen manufacturer. In 2008, the Hefei government and BOE signed a 6th generation TFT-LCD production line project investment framework agreement. In this arrangement, Hefei Construction Investment, Hefei Xincheng, and BOE all put up equity in BOE’s Hefei subsidiary (See diagram below). In China, most corporations have myriad subsidiaries, like BOE, which allows local subsidiaries such as the one in Hefei to take on capital from the local government, while the main corporate entity remained in Beijing. After three years, the government began to withdraw its capital, booking a profit of 700 million yuan.11 Today, the Hefei Jianxiang Investment Co. only owns a bit more than 1% of BOE—mostly it is controlled by Beijing’s state assets supervision group. BOE’s Hefei factories are within the Xinzhan Hi-Tech Area, 新站高新技术产业开发区 (see image and map below).

The Industry Chain for Optoelectronics

Many of the input industries needed for LCD screens have linkages to semiconductor production. For example, Vital Thin Film produces ceramic components that are used in both semiconductors, displays, and photovoltaics. First of all, semiconductors are needed for screens. This includes logic chips to process the signals and data that are displayed on the screen, memory chips, driver ICs to control pixels and display images, and other specialized chips related to display brightness and sensitivity.

Corning branch of American company specializing in glass, ceramics, and related materials and technologies including advanced optics,

Visionox: primarily manufactures organic light-emitting diode (OLED) displays. They produce both PMOLED and AMOLED displays, including those for wearable devices, smartphones, laptops, tablets, and automobiles.

Sunnypol: “Sanlipu Optoelectronics mainly engages in the production and sales of polarizers, protective films, solar films, and optical films; research and development, and sales of optoelectronic materials.” (Qichacha)

See Ya high-tech enterprise specializing in the research and development, design, production, and sale of OLEDoS displays

Nexchip A pure play semiconductor foundry, China’s third largest—a joint venture with Taiwanese company.

Vital Thin Film VTFM’s product portfolio includes high purity metals, complex metal alloys, precious metals and ceramics for Display, Photovoltaic, Semiconductor, Optics, LED, Data Storage and Glass applications.

Supermask 合肥清溢光电有限公司 R&D for design, production and sales of masks. It is one of the earliest and largest mask manufacturers in China, products used for: flat panel display, semiconductor chip, touch, circuit board and other industries.

Read more

Nonetheless, It’s difficult but essential to keep two paradoxical truths about China in the current moment in our heads at one time:

1. China has seen its domestic economy slump over the past five years or so, largely following the severe Covid-19 lockdowns and real estate downturn, and the economy has yet to recover from this.

2. There are several areas of China’s economy in which China has continued to go from strength to strength, particularly manufacturing of electric vehicles, clean energy, and other technologies

Because of China’s sheer size its possible and indeed likely that these two paradoxical situations will continue to coexist for some time, no matter how the Trump tariff war will play out. In this post, I’ll briefly touch on a few dynamics and offer my thoughts on why these matter for policymakers, investors, and anyone interested in “getting China right” (or at least, not getting it entirely wrong) in the near future.

1. Review of Past 5 Years; “How did we get here?”

2. The worst of times

3. The best of times

How did we Get here? Covid 19 and the whiplash of sentiment on China

First of all, its important to review briefly the past five years or so of conventional wisdom on China to see just how often the narrative has flipped. Let’s recall back to the years leading up to Covid. Trump’s first more limited trade war largely failed to put any dent in China’s progress, and Trump was hailed by China’s internet sphere as Chuan Jianguo 川建国 “the founder of the [Chinese] nation.” So during this period before Covid, China’s growth and dynamism was perhaps at its high point. It was also during this period when Xi Jinping began to talk about “changes unseen in generations” 百年大变局 in his view that the East was rising and the West was declining.In 2017, Xi used the term “profound changes unseen in a century”, or bainian wei you zhi dabianju to describe what he viewed dramatic changes in the world order, encompassing technological transformations but also a shift away from a U.S. or Western-led global order.1 As Jin Canrong, Dean of the School of International Relations at Renmin University, put it in 2019, “after the fourth industrial revolution, the productivity of the East is likely to be ahead of the West, or at least a balance between East and the West will be achieved. This is the most important change among the three changes unseen in a century.”2

The pandemic was a watershed moment. At first, the outbreak of the epidemic in Wuhan shattered Chinese faith in the Communist Party’s governance, epitomized by the rare outbreak of anger at the government’s coverup and delay in responding to it. But by the summer of 2020, with the virus successfully contained in China and spreading like wildfire across the rest of the world, China's leadership began to look comparatively good. At the end of 2020, China had come up with a system of containing the virus through strict controls on mobility, the introduction of “health codes” for intercity travel. It was supplying essential goods to the rest of the world. But anger at the rest of the world towards China, fomented in some cases by politicians like Trump, lead to a continued freeze in relations. After Biden’s victory, the two sides began exploring restoring a “floor” of normalcy in what was spiraling into all out cold war. At a March meeting in Anchorage, Alaska, the Chinese delegation lead by Wang Yi and Yang Jiechi excoriated the Americans, telling them that they had no right to criticize China and saying that China had made great strides by controlling the virus.3 This moment was a point at which China’s leaders were feeling emboldened by the control of the virus and the comparative failings evident in the U.S. and other Western nations’ responses.

The Worst of Times?

However as 2021 progressed, the situation in China became less sanguine. The effect of continued lockdowns and travel code restrictions brought down the popular mood and the economy. Xi Jinping was overseeing a crackdown on his own “platform firms” of Alibaba and Didi, which sent shares of Chinese tech giants cratering and brought down investor and consumer confidence. In 2021, a default by home developer Evergrande triggered the property downturn that has continued to this day. Real estate growth, a major economic engine and source of local government revenues, has since dried up.

When I visited China for the first time since Covid in 2023, and again for a longer period in the summer of 2024, I was struck by the pervasiveness of pessimism expressed by almost everyone I talked to (I wrote about my trip last summer, here). In city after city I heard how local governments were out of money, a point backed up by data.4 Housing projects lay unfinished. The value of homes had decreased. A friend in Beijing told me how his family was trying to sell a property in an advantageous location near high-tech parks in Zhongguancun but couldn’t, and even if they could the price they would fetch would be far lower than what the property had been worth a few years back. Artists talked about tightening space for free expression. The dominant word I heard though was not really about the U.S. but about China’s own internal “involution” or neijuan, described to me as a phenomenon of overwork and competition leading to diminished returns for everyone. Sure, people admitted, there were pockets of exceptional dynamism. But these didn’t seem to matter much for most people’s overall financial situation or sentiment about the future.

The Best of Times? Manufacturing and innovation juggernaut

Despite the overall downturn in the property markets and economies, the last five years have been remarkable for the juggernaut of Chinese manufacturing and innovation. The main evidence of this is China’s success in becoming an electric vehicle powerhouse.5 China has rapidly overtaken previous car exporting powerhouses Japan and Germany. While there are basically no Chinese EVs available in the U.S. due to current tariffs on EVs, the EU has become a significant market for Chinese automakers, as are developing countries in South America and Southeast Asia. Criticism has been levied at China’s “overcapacity” in manufacturing, but Trump’s tariffs could ultimately lead to greater Chinese footprint across the world’s major auto markets outside of the U.S. But there is also the situation of “excess competition” as there are around 200 manufactures of EVs in China—many may consolidate or disappear in the coming years.6 And fear of harm to country’s domestic automakers, such as in Europe, may lead to greater restrictions on Chinese EV sales in core international markets.

Implications

It’s worth reflecting on what the bipolarity in China watching sentiment in the West actually means. One the one hand, the extreme views of China say more about the anxiety in the U.S. and West than China itself. China’s rise has challenged the U.S.-led order, and the U.S. itself has seemed to lose faith in the system of free trade it built. China is often portrayed by the media, and by Trump and allies as a threat that the U.S. must defeat in a “Second Cold War.” On the other hand, there is a persistence of the “China is about to collapse” idea pushed for years by those like Gordon Chang and Frank Dikötter who has called China an “empire of illusion.”

However, the bipolarity on China is not only in the eye of the beholder. The current reality in China is in fact paradoxical. The success of manufacturing and innovation in certain high-tech sectors coexists with widespread economic stagnation in the rest of the economy as the hangover from the property boom and bust has not yet abated. The growth of high manufacturing sectors in China has not yet (and may not for some time) lead to growing income or wealth for the majority of the country. My TikTok and Instagram feeds are sometimes littered with people showing videos of China’s cities lit up with evening light shows with captions like “THIS IS A THIRD TIER CITY IN CHINA AND ITS MORE FUTURISTIC THAN ANYTHING IN THE U.S.” Such flashy scenes are seductive, but also aren’t very representative—there are so many small and medium-sized cities in China with widespread unemployment, areas of disinvestment, poverty and regions entirely excluded from the boom in certain manufacturing sectors China is seeing. This is not unlike the geographic divide in the U.S. between left behind regions and superstar cities that many see as fueling the rise of Trumpism.

Having a more balanced view of the current situation in China, both its strengths as well as its continued frailties, is important for policymakers in the U.S. Too often, media and policy discourse alternates between extreme views: China is on the verge of collapse or it is about to overtake the U.S. in every possible way. Understanding where China is ahead (sectors like EVs, manufacturing, materials, rare earths) and where China lags the U.S. (enterprise business services, stock market, trust in its monetary institutions, and lack of appeal as a destination for international immigration and education) would help assuage the current myopic focus in the U.S. on “winning a race of China” where the goals or metrics of victory have never been clearly enunciated. In order to stay competitive, the U.S. can focus on its own strengths and advantages, such as its lead in higher ed and its appeal as a global destination for high-skilled workers. Instead, the Trump administration seems intent on destroying any of those advantages, all in the name of pursuing some romantic and outdated view of American manufacturing from the early 20th century.

In 2017-2018, I spent a little under a year at Peking University’s College of Landscape Architecture. I was helping to set up a collaboration between Harvard Graduate School of Design and the landscape architecture college at Peking University called the “Ecological Urbanism Center” which was going to focus on strategies for retrofitting Chinese cities for climate change and sea level rise. The collaboration itself was short-lived, cut short by disagreements between the university departments over funding (perhaps a harbinger of the broader deterioration of the U.S-China relationship)—but that’s a story for another day!

While based at the beautiful campus of Peking University I got to know a group of energetic and creative students of landscape architecture. The department is headed by the well-known landscape architect Yu Kongjian 俞孔坚, who founded the large firm Turenscape 土人设计, and has designed numerous flood-resilient sponge-city projects around the country and worldwide. 1 During their time as students at PKU, they built a small demonstration “rainwater garden” on the site of the college, adjacent to the Weiming Lake. This small project would eventually lay the groundwork for their practice.

In the summer of 2024 I caught up two of these students in Shenzhen, Huang Binling and Yuan Zhenyu. I hadn’t seen them since 2018 when they were still master’s students. Six years had passed, and two of them had move to Shenzhen and in 2019 founded a landscape architecture practice called Sketch Action 草图营造.

The two friends had been invited during Covid to move to an urban village in Shenzhen and set up their practice, and they had since made the village both their professional testbed as well as home base for them and their budding team of nearly 30 or so mostly youngish employees. Sketch Action’s practice is interdisciplinary, drawing on landscape design as well as participatory community-based planning. Sketch Action works alongside residents to co-create new spaces and help residents take a leading role in shaping and improving their own communities. The practice has also designed facilities for child-friendly and ageing-friendly communities throughout Shenzhen.

After an almost hour taxi ride from Shenzhen’s rail station I arrived at the Shangwei Arts Village 上围艺术村, located in the northern Longhua District 龙华区 of Shenzhen. There are 372 urban villages in Longhua District alone. Shangwei was an unassuming and modest village, ringed by small warehouses and repair shops. The village itself had been designated an arts and cultural village, and was home to a motley group of artists, sculptors, and other creatives.

The practice is dedicated to “community-gardening” as a form of regeneration of urban villages 城中村. Urban villages are a uniquely Chinese phenomenon, stemming from the confluence of rapid urbanization clashing with the existing dual land structure of Chinese villages/cities—village land is still classed as 集体 “collective”, with each villager entitled to collective ownership, often entailing rights in communal enterprises and right to build a house and farm. When Shenzhen urbanized, villages like Shangwei and many others were often able to keep their residential property, while farmland around the village was absorbed for urban development. Today, Shenzhen’s urban villages provide low-cost housing, vibrant street life and food. While many Chinese cities have been moving to demolish their urban villages, many of Shenzhen’s villages have become quite wealthy, undertaking their own property development and renovating homes, maintaining their “right to the city”.2 Villages in southern China, such as in Shenzhen and Guangzhou often have strong lineage organizations and corporatism, making them formidable forces in local politics.

I was greeted in the courtyard of the compound, which houses the studios, living space for the founders and some of the other architects, kitchen spaces. A video playing on the loop in the courtyard features Huang talking about his journey to Shenzhen and the idea behind the practice. I gave a presentation on my own research, followed by a conversation with the two founders and some of their staff.

Despite a downturn in the real estate and construction sector in China, Sketch Action had found a niche in small-scale village renovation projects. While each project is not hugely profitable (around 150,000 RMB, or $20k USD), they can do many small projects quickly, building up experience and scale. They have already completed over 100 community renovation projects in Shenzhen. The two young founders have already established a name for themselves in only a few years after graduating. As they told me, “we’re doing a small-scale approach so we are relatively lucky, this type of small scale revitalization is actually now taking off whereas the large-scale projects are slowing down.”

Huang and Yuan promote their approach as “community-led design”, borrowing from ideas about bottom-up urban planning and community governance in the West. They argue that their practice is “as much about governance 治理 as about design 设计. Although, even in the projects they have worked in, the commissioning entity is often the local village administrative committee, which is technically part of the state apparatus. Thus, the question of how much “grassroots governance” is really possible within China’s hierarchical administrative system is an open one. Huang and Yuan acknowledge this, but also state that they are commissioned by the local village authorities to undertake community planning, working with residents in the village on behalf of the local government, which lacks the capacity or understanding of community-driven design. So in this sense, they are consultants hired by the local village government to interface and work with the local population.

Their projects also span typologies from community gardening to public services. One of their projects in Shenzhen’s Window of the World neighborhood was built in a community with many foreign residents from 22 countries. Here, residents were selected by lottery to obtain small plots for farming vegetables, “including shared planting areas, gravel gardens, four-season orchards, herb gardens, bee and butterfly gardens, rainwater gardens, and activity squares.”3 At another project in the Huaiyayuan Community of Nanshan District, Sketch Action helped a community transform an empty paved area of a parking lot in front of the community into a diverse garden, and residents participated in planting and design themselves.4

In recent years, architecture and planning profession in Chinese cities have been discussing and promoting notions of “micro regeneration” or wei gengxin 微更新. Influenced by community-driven design in Japan, Europe, and elsewhere, the trend towards micro renovation comes on the heels of China’s massive urbanization, which has seen massive transformation of urban areas on a macro scale. As urbanization and growth slows, transforming urban neighborhoods through smaller-scale renovation is a more financially feasible and less disruptive approach to urban redevelopment than the wholesale demolition or chaiqian 拆迁 that characterized much of China’s rapid urbanization in recent decades.

I also spoke with Huang and Yuan about how their practice fits within evolving dynamics of local governance in China. Many media reports on China suggest the country has become more authoritarian during Xi Jinping’s presidency. And while Xi has undoubtedly centralized power at the central/national levels, at the grassroots level the story may be different. One thing the founders told me stuck with me—""the state increasingly wants local citizens to manage their own affairs if they can, the state wants this kind of grassroots appeal.” This also has to be seen against the reality of China’s local governments being on precarious financial footing since the ongoing real estate downturn drained a key source of revenue—land leasing. Local governments have limited funds for large-scale village demolition and reconstruction now, but small-scale community beautification efforts, such as the ones advocated by Sketch Action, are more in vogue now.

Do such small-scale efforts like the ones pursued by Sketch Action offer a future for China’s neglected urban villages and emptying rural areas? On the one hand, who can argue with the efforts to beautify urban villages through greenery and bottom-up design? To be sure, Shenzhen is one of the more progressive and liberal Chinese cities, so this too must be acknowledged. Sketch Action has worked in Wuhan but they acknowledged city government in Shenzhen is more open to these sorts of approaches than elsewhere. And of course, on their own such small-scale efforts will not be able to solve more intractable problems such as the emptying of the countryside, youth unemployment, or other problems of rural underdevelopment in China.5 But the visit to Shangwei Village and my conversation with Huang and Yuan were inspiring, that despite the fact of increasing centralization and authoritarianism at China’s top levels, there are counter-trends and alternative possibilities blooming at the local level.

I rebranded my substack as Sinocities in June 2024, with this welcome post noting that,

“China is such a vast country that any attempt to grapple with its complexity needs to understand regional and local particularities. There’s a lot of good China content in the English speaking world, both from international scholars but also increasingly China-based scholars writing in English. But there’s still very little that focuses on specific locales and specific places as a lens into the complexity and variety of China’s development.

Since then, I’ve been pleasantly surprised and humbled by your interest in my work.In general I try to illuminate somewhat obscure topics that have not received much attention in English media (or Chinese media), such as the building of China’s nationwide data network, Xi Jinping’s construction of a new Archives 版本馆, or heritage tourism such as in Shanxi province. I’m not just interested in esoterica for esoterica’s sake. I also try to explain why these often overlooked and understudied topics actually matter—such as the importance of data centers to China’s digital infrastructure, or the importance of ideology and historic memory to contemporary governance under Xi. And throughout all of this, there is a focus on place and specificity—bringing under-examined places and stories to light, and examining their implications for broader issues and themes.

For the most part I intend to keep on doing what I’ve been doing. But I also have realized that the amount of time and research that I put into some of these stories is not without a significant time cost. The time in takes to undertake the research goes beyond the process of sitting down to write, and involves pulling in various types of data from geospatial to news reports from Chinese media, scholarly research, and interviews. Balancing this alongside ongoing research and teaching commitments at MIT, where I currently work, requires dedicated planning and time.This is why I have decided to turn on subscriptions and allow a paid subscription model.

However, to be clear, I intend to maintain a significant portion of Sinocities as free and open to everyone. For example, I intend to keep posts like CITY REPORTs (such as the one I’ve done recently on Chengdu), as free for all. For some in-depth reports that require a lot of research and data collection, I may paywall portions of the story, allowing free subscribers to get a short overview of the topic, while reserving in-depth reports for paying subscribers. The intention is not to put all of my content behind a paywall, but offer a way for dedicated subscribers who feel they benefit from the unique stories and perspective I bring to contribute to my work. For those of you who value and support my work, please consider subscribing. I will also think about developing unique and customized reports for paying subscribers—an idea that is still in development.

Subscribe now

For those who do not have the resources or time to purchase a paid subscription, I value your engagement and continued interest and I will continue making a good deal of what I write and research available for non-paying subscribers. Thanks to everyone for your interest and support of my work, and please feel free to reach out with suggestions, ideas for topics or places you’d like to see covered, or any other feedback!

I’ve written before about the Eastern Data Western Compute 东数西算 project, China’s effort to boost its’ “computing power” by constructing new data centers in the country’s West. So far, I’ve mostly focused on efforts to build out new clusters of data centers in western locations including: Zhangjiakou (Hebei), Hohhot (Inner Mongolia), Gui’an New Area (Guizhou), and Qingyang (Gansu). But the EDWC project is, at least in theory, more than the sum of its parts.

The vision of the EDWC is not only about building new “nodes” of data centers, it is also about creating a “national network of computing”, 全国算力一张网 quanguo suanli yizhang wang so that computing power can be shifted between data centers depending on fluctuations in computing demand and supply in different regions, not unlike the way interconnected or smart power grids can move electricity around the grid depending on where demand is greatest. One of the premises of the EDWC is that data centers, which consume large amounts of energy, should be located in areas with ample renewable energy supply and cheaper land and energy than in the populated east coast. Networking computing resources across the country can make better use of energy in the West without having to transmit electricity from West to East.

So far, the “computing network” aspect of the EDWC plan has received very little attention even though it may be one of the more significant aspects of the plan. I’ve not written about it before because details are hard to come by, and the implementation is not clear, even though some of the conceptual and technical ideas have been laid out in research papers by China Telecom, Huawei, and the CAICT—an influential think tank of China’s Ministry of Information and Internet Technology (MIIT). But I will try to detail here what we know so far, and what has and is being done on this:

“The EDWC project is not only about building new “nodes” of data centers, it is also about creating a “national network of computing”, 全国算力一张网 quanguo suanli yizhang wang so that computing power can be shifted between data centers depending on fluctuations in computing demand”

A National Network of Computing Centers

In a televised interview in 2022, Yu Xiaohui 余晓晖 the President of the influential think tank CAICT1, which has played a leading role in developing the EDWC plan, notes that “other countries are doing similar things [to the EDWC] but only within one cloud company, but this is an advantage of China, we are doing a systematic layout.”2 In other words, what he alludes to is the fact that the EDWC project is a state-led effort to coordinate the data center infrastructure of the entire country, which requires going beyond simply encouraging cloud/telecom providers to invest in their own new cloud computing centers. It also means creating a unified national computing network that would allow for a more dynamic allocation of computing demand across the country.

But to do this requires creating a system that can allocate demand from computing centers of different providers. From a business and data privacy standpoint this seems difficult to do. However, China’s three state-owned telecom operators are the ones playing a significant role in building out the EDWC project. In a 2023 research article, engineers from China Telecom outline the basic idea of what they term “computing power gateway” 算力网关:

“The computing power network is a new network technology that distributes computing, storage, algorithm and other resource information of computing power nodes through the network, and combines network information and user needs to provide the best distribution, association, transaction and allocation of computing, storage, network and other resources, thereby realizing the optimal configuration and use of various resources.”

“As the core network element device in the computing power network, the computing power gateway is based on computing power measurement and computing power identification, and transmits and publishes related computing power strategies and data forwarding through core functions such as computing power routing and computing power perception. It is the basis for realizing the integrated scheduling of computing power network”3

Dispatching Centers 调度中心

In various documents and news announcements there have been reference to 调度中心 or “adjustment centers” in each of the data center nodes, which are supposed to function as traffic centers for nationwide computing resources, balancing supply (western data centers) and demand (eastern applications). Such “adjustment centers” allow computing tasks to be dynamically allocated to different data centers, such as allocating workloads between 8 national computing hubs and 10 data center clusters, optimizing latency, and improving energy efficiency by redirecting “non-urgent” data tasks to western renewable-powered hubs during off-peak hours. A patent application filed in 2023 by a Beijing-based company CASDAO (算力互联) outlines the basic requirements for such a “data center scheduling center”, using AI to assign compute requests to different data centers based on matching the requirements (i.e. latency) of the specific task with a particular data center based on availability and useage levels.4

In Suzhou, China Telecom helped open a “dispatching center” for the Yangtze River Delta Region. A scheduling center for the Greater Bay Area was opened in Shaoguan, the EDWC hub for the region, connecting five supercomputing, intelligent computing, and cloud computing centers, and housing a“scheduling center” for different companies under one roof (image below, right). Most of the “dispatching centers” I could find evidence for in news reports seem to be connecting data centers within metropolitan regions, such as the Greater Bay Area or the Yangtze River Delta, and serving as gateways between these regions and the inland cloud hubs being constructed in the West.

Private firms have also been developing their own cloud network solutions to optimize network efficiency. Huawei outlines some of the technical innovations they are proposing to thicken network connectivity between East and West, such as: i.) “intelligent algorithms for data center network switches,” and ii.) “an Intelligent cloud-map algorithm” 5 So far these innovations seem to be deployed within Huawei data centers, but its not clear to what degree they can be deployed to strengthen connections between data centers of different operators

Obstacles

There are technical challenges to developing the national network. But there are also obvious financial/proprietary hurdles as well, namely how to interconnect cloud networks of separate companies who have their own proprietary businesses and systems. This may be easier to do with the three national operators (China Mobile, China Telecom, and China Unicom) than it is with private cloud operators, which are still the leading cloud platform companies (Alibaba, Tencent, Huawei, Baidu).

As researchers noted in 2023, “Although the compatibility of heterogeneous cloud resources is technically easy to achieve, it is not in the interests of proprietary cloud vendors and may be difficult to promote in the short term.”6 In the U.S., there are some examples of so-called “multi-cloud” interconnections, such as agreements between cloud operators Microsoft and Oracle that allow clients to connect certain cloud databases stored on different cloud platforms.7 But China’s ambition to create a national computing network would require a much greater interconnection and coordination to carry out. The degree to which China is able to realize this network remains unclear. Some efforts have been made to connect data centers within regions, but this does not yet appear to have occurred on a large scale, or with the operators of China’s leading cloud platforms.

Implications

“Networking of data centers in the EDWC plan has implications for the ability of Chinese firms to develop advanced AI models and services, leveraging new computing centers being developed in Western China.”

China’s efforts to link up cloud facilities has also received some attention in the context of its efforts to find workarounds to U.S. export controls on the most advanced semiconductors. Reports last year suggested many Chinese firms were accessing advanced chips to train and run AI models via cloud services, some of the operated by Western companies.8 This could potentially happen in Chinese-owned data centers within the U.S. (Alibaba and China Mobile have a few), although that is unlikely given China’s own data export restrictions.9 Recent reporting by the Wall Street Journal suggested data centers in Southeast Asia, some of them run by Chinese cloud operators like Bridge Data, are also serving as hubs for AI training by Chinese companies.

The release of DeepSeek’s model also revealed the degree to which Chinese AI companies are finding ways to train AI models with less advanced chips. Some of this involves software and process innovations, and some of this has involved networking and linking of distributed data centers. As researchers from the Pengcheng Laboratory in Shenzhen noted in a 2023 paper:

As the scale of model parameters continues to grow, the computational resources required for model training become significantly larger. This often leads to situations where a single computing cluster is insufficient to meet the training needs of large-scale language models. Cross-cloud joint training of large-scale language models has emerged as an effective solution to addressing this challenge.10

Therefore, the networking of data centers in the EDWC plan has implications for the ability of Chinese firms to develop advanced AI models and services, leveraging new computing centers being developed in Western China.

Networking previously isolated cloud systems offers potential benefits including:

1. greater efficiency of cloud usage—balancing supply and demand for cloud computing across and within regions; this could help reduce traffic and slowdowns due to overuse of cloud services, and minimize wasting of computing power and energy

2. Pooling of different cloud resources—allowing clients to take advantage of multiple services offered by various cloud providers (this is known in the U.S. and elsewhere as “multi-cloud” strategy)

The degree to which China is able to build out the national network of cloud computing will have implications for its digital innovation, particularly in AI. While the U.S. leads China in the number of data centers by a wide margin, China’s system could develop in ways that diverge from the proprietary open-cloud model in the U.S. in which large enterprise cloud platforms dominate the market (AWS, Microsoft Azure, Google Cloud). Whether and to what degree China’s existing cloud providers continue to dominate the market will depend on their ability to innovate and maintain their edge in the face of increasing entry by state-owned cloud providers into cloud markets. Most likely, China’s state-owned telecoms will continue and strengthen their role as operators of critical network infrastructure, including the network described in this article, while private cloud providers continue to dominate consumer and private-enterprise facing cloud applications. But if China is able to create a more interconnected cloud network, it could herald the arrival of a unique integrated cloud infrastructure for the whole country.

Whereas previous new areas sought to attract private and foreign investment, today most of the new areas I will detail here are anchored by state-owned enterprises and national research institutes.

Introduction

China’s urbanization and economic growth over the past several decades has been defined by the creation of special economic zones (SEZs), new districts, new areas, and other variations on the theme. Perhaps it is no surprise that even amidst a real estate downturn and general economic malaise gripping China at the moment, there are still a number of key infrastructure projects going forward. Among these are several new areas that are developing apace, defying the naysayers and almost every economist out there who has warned that there is little demand for more real estate and that China’s investment-centric model needs to change course. These “new type new areas” developed in the last few years are focused on innovation and scientific research, as well as on promoting “coordinated development” 协调区域发展 between neighboring jurisdictions. Whether there is actual innovation or “coordinated development’ happening in these places is another question, but that’s at least how they are being promoted.

In general, new areas under construction now are not dramatically different to new areas planned over the past few decades: large greenfield sites, laid out en masse on the edge of cities with overly wide avenues and out of scale buildings that aim to attract key industries and create new ‘growth poles’ of urban development in their cities and regions. But at a time where the Party-State is commanding a greater presence in China’s economy, unicorns have gone missing, and the urban middle class is not spending like they did before, the chief anchors of new area construction in recent years are different. Whereas previous new areas sought primarily to attract private and foreign investment, today most of the new areas I will detail here are anchored heavily by state-owned enterprises and national research institutes. Even as private capital may be hesitant to invest, state-owned institutions have been the first to move into many such zones.

Of course the “state” has always been involved in creation of special economic zones particularly after Shenzhen, and the following 19 national new areas. But over the last few years we see a pattern in which new industrial districts are no longer as focused on attracting foreign or domestic private investment but are a.) primarily anchored by central state institutions. These include state research institutes, data centers of major telecom operators, and new research bases for central state-owned enterprises and other elements of “State capital” more generally. Much of the residential real estate is also built by property arms of central SOEs (China Rail, China Communications), suggesting central SOEs are playing a bigger role investing in such state-prioritized projects.

China’s new areas have long been heavily supported by the local and national state, but the firms and tenants in emerging districts appears to have a distinctly state-owned character. This is not all that surprising, given the times, but it’s worth exploring a bit a few recent new districts under construction across China to get a feel for what industries and sectors have received central support amidst a trying economic slowdown, and how this takes shape in the urban (and suburban) environment. There are also some design features that can be seen across these districts:, b.) ecological elements are favored, such as “sponge city” design—permeable parks, lots of wetlands, and in general an emphasis on integrating buildings with nature—this is not a bad thing! Restrained architecture is favored over eye-catching starchitecture and there is a focus on spaces for reading (libraries, book stores, etc). Finally, from a governance perspective some (but not all) of the areas are coordinated efforts of neighboring jurisdictions to coordinate development across city regions and create city clusters.

Let’s hop on a whirlwind tour of some of China’s emerging new areas…

Xiong’an New Area 雄安新区

I won’t detail too much here on Xiong’an, since I’ve already written extensively on it for Sinocities here and here. While many people have already declared Xiong’an a failure or a ghost city, construction continues—primarily driven by state capital and the relocation of several large central enterprises. The city initially grew out of the 2015 Jing-Jin Ji plan for integrated development of Beijing, Tianjin, and Hebei. Only in 2017 was Xiong’an announced as a major pillar of the regional effort. As part of this, Beijing and Tianjin-based companies, investment funds, as well as a great deal of central government support have been coaxed to support Xiong’an’s development. Four branches of Beijing-based universities are also opening up satellite campuses in the area. After nearly eight years of development, there are many uncertainties facing the area, and skepticism over the merits of trying to develop such a city in a location with few natural advantages—essentially an entirely state-driven project. However, that’s exactly how to view Xiong’an—a state-driven effort to relocate SOEs to create a new cluster of state-led innovation in renewable energy, infrastructure, and other national-security critical tech, bolstered by several branches of Beijing universities—albeit crucially none of the nation’s top universities like Beida and Tsinghua.

Companies with Significant Investment

• Huaneng 华能

• Datang 大唐

• China Telecom 中国电信

• China Satellite Communications 中国星网

• State Grid 国家电网 (see above)

• China Communications Construction 中交集团

• China Rail Construction 中铁建

• Sinochem 中华

• Power China 中国电建

• China Energy Engineering Corporation 中国能建

Universities

• Beijing Forestry University 北京林业大学

• Beijing Geological University 北京地质大学

• Beijing Technology University 北京科技大学

• Beijing Jiaotong University 北京交通大学

Yangtze River Integrated Development Model Zone 长江三角一体化模范区

Straddling three province-level jurisdictions of Zhejiang, Jiangsu, and Shanghai, the Yangtze River Coordinated Model Zone aims to develop the region lying at the interstices of three of China’s most developed regions. To see more details about this project, see this earlier post I did on the area. Key tenants include a large research campus of Huawei, a data-center cluster of the East-West Compute Plan. The development is an effort to promote cross-provincial coordination in planning, responding to top-down policies of “Integrated regional development” 协同区域发展promoted at least since 2014 when it was included in the National New Type Urbanization Plan. In the past, Shanghai and Suzhou have each developed with their own projects—Suzhou’s vast science/research areas, and Shanghai’s Pudong. This is the first real effort to coordinate growth between what are some of the most economically cities in China.

Chengdu Tianfu Science City 成都天府新区 （科学城）

Lying at the southernmost terminus of Chengdu’s Metro Line 1. which runs North to South through the center of the city, Science City (科学城） is the newest area under development in the sprawling Tianfu New Area, which was inaugurated back in 2011. Already completed are branches of the China Academy of Sciences, a National Cloud Supercomputing Center, a research institute of the Sichuan-Tibet Railway, and office parks developed by China Communications 中交集团 and Power China 电建.

other tenants in Tianfu Science City :

• Southwest Geological Technology Innovation Center西南地质科技创新中心

• China State Construction Engineering Green Building Engineering Research Center中国建筑绿色建造工程研究中心

• Tianfu Digital Smart Industrial Park 天府数智产业园

• China Academy of Sciences Chengdu branch 中国科学院成都分院

• China Academy of Sciences Biological Research Chengdu 中国科学院成都生物研究所

• 卫星通导及系统融合产业园 (CCDI悉地国际受邀为国家级高新技术企业（成都盟升电子集团）)

• Chengdu Western Electric Network Safety Research Institute成都西电网络安全研究院

• Power China 中国电建金结研发中心

• Tsinghua Sichuan Energy Internet Research Industry Ctr清华四川能源互联网研究院产业生态中心

Xi’an Xixian New Area 西咸新区

Xi’an is the rapidly growing Western metropolis and capital of Sha’anxi Province. The neighboring city of Xianyang, the location of the ancient capital city of the Qin Dynasty, has lagged behind Xi’an in development. The new Xixian New Area was made a national-level new area in 2014 and straddles both cities, thus facilitating the creation of a Greater Xi’an Metropolitan area. There are several sub-areas within the sprawling 882 sq km area, including an airport-related development around Xi’an-Xianyang Airport; Qin-Han New City—a real estate development located near the old Qin and Han imperial city site—and also the site of a new Sha’anxi History Museum Qin and Han Pavillion. The Western Cloud Valley 西部云谷 contains cloud computing centers and other tech and innovation-focused companies. On a visit in August of 2024, the area was fairly quiet and did not appear to be particularly well occupied. But some sections of the area were already built out and partially occupied.

• China Telecom Data Center 中国电信

• China Unicom Data Center 中国联通

• China Mobile Data Center 中国移动

• Shanxi Data Group 陕数集团

• China Railway Construction NW China HQ 中铁建工西北总部

It’s lunchtime at the Xinghe COCO Park shopping mall in the Longhua District 龙华区 of Shenzhen. I’m meeting with a technologist who has also been working as a planner for Shenzhen’s “low altitude economy” or dikong jingji 低空经济. He runs a small consulting firm with an office in a building above the mall where he shows me a plan for the low-altitude economy he’s working on for the district. A map outlines the flight routes 航线, operated by delivery companies like Meituan 美团 and Shunfeng 顺丰.

In 2024, the NDRC (National Development and Reform Commission) of China established a unit focused on the low-altitude economy responsible for formulating mid- and long-term development plans and policies related to the low altitude economy. This refers to drones, drone delivery, and unmanned aerial vehicles under development such as electric vertical take-off and landing (eVTOL), operating under 1000 m altitude.

He takes me out to a food delivery stand operated by food delivery company Meituan, located in an outdoor plaza of the mall. Although it’s lunchtime, there are no drones operating. Puzzled, we find a sign on the stand saying “drone delivery currently paused citywide. A bit disappointed, my host tells me, “lunch hour is usually the peak.” As we walk back to his nearby office, we pass office workers grabbing their lunches from delivery boxes located nearby—presumably delivered the “old fashion way” by delivery courier on e bike.

I speculate (without reply) if Xi Jinping’s visit on the same day to nearby Macau is one reason why drones might not be flying today. But no matter the disappointment on not being able to see drones buzzing overhead delivering lunchboxes, I am later shown a variety of plans and guidelines currently under development for the planning of the low altitude economy in Shenzhen.

Why Shenzhen, and Why Longhua?

Shenzhen has become known as 无人机之都 “the capital of drones”. A big reason for this is the city’s vibrant tech and electronics ecosystem. Founded in 2006, DJI (Dajiang) 大疆 is the world’s leading producer of commercial drones and its HQ in nearby Nanshan District. Meituan, a leading food delivery platform, also has its headquarters in Shenzhen not far from the mall where we are visiting.

“Why Longhua? This is one of two main districts, along with Bao’an (where Shenzhen airport is) where drone delivery is piloting now. This area is a shuniu, a hub. We have Shenzhen North Station.” He points to a map showing delivery hubs, as well as red lines and nodes of pickup stations spread around the district. “Several areas are basically not allowed to fly, such as central urban areas, areas around sensitive infrastructure. Nearby is a small airport for private jets that may be repurposed for logistics and drone delivery in the future.”

One bureaucratic obstacle to the development of the low altitude economy in China is the military. “Much of the airspace in China is controlled by the military. So city governments have not had a lot of autonomy to develop. However, once the higher ups give approval, companies can begin operating in select areas,” my host tells me. The Chinese government recently gave six cities the right to explore the low altitude economy up to 600 meters or lower including Shenzhen.

Government-Enterprise Partnership

Why has China surged ahead in drone delivery? I asked my host. He mentions a slew of technical standards and research from the U.S. But primarily, he tells me its about commercial factors and scale. “Really its being led by industry here. The government didn’t know anything about this initially. But once the delivery companies like Meituan started to explore this, the government starts to follow with legal recommendations, guidelines, etc.” In Shenzhen, because of the presence of delivery companies like Meituan and drone-makers like Dajiang (DJI), there was a natural ecosystem for the development of the industry. And the city does not face as stringent of restrictions on its airspace from the military, compared to cities like Chengdu, which is also trying to promote the low-altitude economy.

Security Obstacles

Bureaucratic Obstacles

There are a number of different government entities that have varying degrees of authority over aspects of the low-altitude economy. This includes first and foremost the military. But also local transportation bureaus, the public safety bureau, and city planning agencies are all required for various aspects of the low-altitude economy. The city is working on planning air ways or 航线, which requires coordination with higher-level authorities particularly the military, and transport bureau. Then, there has to be market demand and a density of delivery spots to make drone delivery possible.

Next steps for China’s Low Altitude Economy

Drone delivery in Shenzhen is just one of the scenarios that might comprise an expanding low altitude economy in Chinese cities. With Chinese car companies piloting flying cars of various kinds, the future of the low altitude economy in China may just be getting started. If it scales up, this will bring a host of critical planning and logistical questions that could fundamentally transform cities and the way they operate—how do you restrict air rights or trade off between building rights and the need for delivery? How would mid-air transport corridors change real estate markets? For example, upper floors of buildings would become more accessible, potentially leading to more commercial or recreational facilities in upper levels of structures. The demographics of China’s shrinking population and housing downturn could mean demand for new construction and density wanes over the next few decades, making some of the planning for the low altitude economy somewhat redundant. But given the degree of interest, the companies developing flying car protoypes and drones, and central investment into the low altitude economy, we likely are just seeing the beginning of what could be a major Chinese urban innovation.

The Rise of Chengdu? China’s Western powerhouse…

Chengdu has a reputation for being a relaxed fun-loving city, with residents whiling away their weekends chatting and playing cards in teahouses. Oh, and did I mention pandas? If you forgot Chengdu was the home of giant pandas, you will have plenty of opportunities in the city to be reminded— giant stuffed mascots are plopped in front of trendy stores, and many tourist stands sell nothing but Panda plushies of various sizes and positions. If the pandas knew how much revenue the city made from their image they would surely demand royalties—but they’re too busy sleeping and eating bamboo all day.

Of course, Chengdu is not all teahouses and pandas. It’s the 4th largest city in China and has become an economic engine of Western China and a contender to become the first “first tier” city in inland China. What is driving the rise of Chengdu? A recent Economist article highlighted it in a short report on inland Chinese cities—they’re cheaper and have a more relaxed pace of life than in Beijing, Shanghai, or Shenzhen. 1 The “hibiscus city” 蓉城 (a nickname of Chengdu) has long been known as a socially tolerant and open city—with one of China’s hottest gay scenes (earning the nickname Gay Du). So the city’s quality of life is not unrelated to its economic attraction—amidst the prioritization of human capital and creative industries, urban quality of life is a major competitive advantage.2 It’s something that puts Chengdu ahead of Zhengzhou and probably Wuhan in the race for attracting talent among inland cities. But “soft power” alone also doesn’t fully explain the rise of Chengdu.

Don’t be fooled by cuddly pandas and buzzing gay bars, Chengdu also owes its economic rise partly to its role as a hub for defense industries and central state capital in Western China.

Don’t be fooled by cuddly pandas and buzzing gay bars, Chengdu also owes its economic rise to its role as a hub for defense industries and central state capital in Western China. Chengdu is the HQ of the PLA’s Western Theater Command, formerly the Chengdu Military Region. Chengdu Aircraft Industry Group (CAIG) is one of China’s leading military aircraft manufacturers, and currently developing some of China’s next-gen stealth fighter jets.3 A businessmen I spoke with recently working on the “low-altitude economy” (drones and autonomous flying vehicles) noted the tight restrictions of the military on airspace around Chengdu as a key obstacle to growth in this emerging sector. But the importance of Chengdu to the military and defense means the city is a major focal point of central investment.

As a recent article writes, “With a comfortable coat on the outside and a fighting spirit on the inside, Chengdu has always had two sides: the city of comfort and the city of striving. Just like the two sides of Chengdu, comfort is the high-profile A side, and fighting is the low-key B side. And all the comfort comes from fighting.”4

Compared to the also bustling metropolis of Chongqing, Chengdu is the political and cultural capital of the region, whereas Chongqing has long been a riverine port, manufacturing hub, and logistics center. Chongqing was separated from Sichuan in 1997 to create the fourth “directly ruled municipality” or 直辖市, making Chongqing functionally equivalent to a province. The more direct comparison for Chengdu would be Xi’an: a historic capital, a provincial capital, and a center of universities and defense-related industries. Chongqing may have also fallen out of favor during the Xi Jinping years due to its association with Xi’s former rival Bo Xilai 薄熙来, the former Party Secretary of Chongqing who built up a power base in the Southwestern city before being sacked and arrested in a high-profile scandal that paved the way for Xi’s rise to power in 2012. When asking a Chengdu local to compare the two cities, he told me, “Chongqing hot pot is spicier than Chengdu hot pot, the people are spicier, and the landscape is more dramatic with mountains and narrow buildings. Chengdu people are more 温和 wenhe- (mild).” Despite this competition between the two, official policy is now aiming for the “Chengdu-Chongqing Economic Circle” 成渝城市圈 to more closely integrate the two into a 城市群, on par with the three coastal city clusters (JingJin Ji, Changsan Jiao, Zhusanjiao). One example of this is Chengdu’s new Tianfu International Airport and airport city, located far to the Southwest of the city,

Chengdu Teahouses as a “third space” and “public sphere”?

I hadn’t been to Chengdu since 2009, the first time I ever came to China as an exchange student. One place I wanted to revisit was People’s Park 人民公园 in the center of the city with its famous lakefront teahouse, Heming Tea House 鹤鸣茶社. While Chengdu had transformed in the 15 years since I first saw it, this famous “cultural landscape” had hardly changed at all. On a warm Saturday it was thronged with locals drinking tea, playing cards, and gossiping. I couldn’t find a table in the main lakefront area so I had to pay 50RMB to watch a Sichuan opera performance in a sectioned off area. People’s Park Teahouse is a prime example of the kinds of “third places”5 that are increasingly disappearing in cities around the world due to the rise of teleworking, homogenization of malls, and the decline in socializing among Gen Z youth (particularly in the U.S.) Chengdu’s teahouse seemed to attract all ages—the young, old, families alike.

Chengdu: Legacy of the Planned Economy

Like many Chinese cities, the 1950s and 60s saw a reorganization of urban space to suit the ideology and needs of the planned economy. Outside the old city, factory districts were developed on the periphery of Chengdu—a pattern also observed in many other historic Chinese cities like Xi’an. A 1989 CIA map of Chengdu shows major industries in Chengdu (orange) as well as core urban areas (grey), and educational and other public facilities (purple). The map gives an idea of where some of the main industrial areas were located (in the Northeast of the city). According to Julie Miao, the most important factories located in the Northeast of the city due to favorable winds, and higher elevation.6 Factories such as the state-run Hongguang Electronic Tube Factory that made cathode ray tubes for TVs and other electronic devices were some of the city’s important industrial anchors. Today, one of these factories was turned into a new cultural space: Eastern Suburb Memory 东郊记忆—Chengdu’s version of Beijing’s 798. Today, Chengdu is a much more diversified economy. The Hi-Tech Zone was set up in 1988 as one of the first in the country. Chengdu is also the largest financial center in Western China. It has also attracted significant foreign investment over the years, with companies like IBM, Samsung, and SAP. As recently as 2017, IBM was trying to open a semiconductor factory in Chengdu7, although IBM recently announced plans to layoff and downsize its research and investment in China.8

COSMO: Youth Fashion

While I’m hardly a fashionista, a friend recommended me to visit COSMO to see some unique local brands. I enjoyed it much more than visiting a generic mall, as most of the shops were selling domestic brands you wouldn’t find elsewhere.

Hotpot

No trip to Chengdu would be complete without hotpot. A local friend recommended this place: Wuliguan Hotpot 五里关火锅, located in a residential area just north of the main shopping area around Chunxi Lu 春熙路. The area itself was interesting to explore as a small-scale down to earth neighborhood of restaurants, bars, and small shops mostly occupying old housing blocks. This area is within the northern walls of the former old city. The hotpot was spicy, but the best part for me was the oil—served in small canisters that you grabbed at the spice station. No foreigners seen here except me—always a good sign :-).

New Chengdu: Tianfu New Area and Science City

Chengdu had grown significantly since the first time I had visited in 2009. Much of this growth has taken place in the suburbs, and particularly in a narrow corridor stretching South of the city center. The Chengdu High Tech Industrial Development Zone was opened in 1988 as one of China’s earliest high-tech parks. Further to the south of this, the Tianfu New Area 天府新区 was incorporated in 2011. Many of Chengdu’s’ new industrial and office parks and high-end housing developments have clustered along this north-south corridor along subway lines 1 and 18. Chengdu Tianfu Airport opened as the city’s second international airport, and is located to the SW of the city, an hour subway ride from Chengdu South Railway Station. I briefly visited the Chengdu Science City 成都科学城 development which is located at the far southern end of Subway Line 1. I was particularly interested in seeing a dazzling “Supercomputing Data Center”, although I was quickly told to leave by security guards after ambling inside the landscaped pathway that wound through it. Hey I just thought it was a park, take it easy…Unlike the earlier developments just to its north, the Science City Area is almost entirely dominated by state capital—research institutes of China Railway, Chinese Academy of Sciences, and other SOEs, and housing developments built by state-owned enterprises China Communications and China Rail Construction.

Despite China’s real estate downturn, Chengdu remains a desirable location in China. The quality of life makes it attractive for liberal-minded youth from coastal cities who want a cheaper cost of living without sacrificing culture, amenities, and high-tech jobs. Chengdu Science City has received significant investment from major central enterprises and research institutes—a branch of the China Academy of sciences is located here, along with a research institute of the China-Tibet Railway and other SOE projects.

Guiyang

I had been following the development of southwestern Guizhou into a cloud computing center in Southwest China for some time. This January, I finally got to visit for a quick trip. Through pre-arranged contacts with several academics at local universities, I was able to get a tour of both Gui’an New Area, where most of the data center construction is, as well as learn a bit more about some of the interesting architectural and urban elements of Guiyang, the provincial capital.

The mountainous province of Guizhou has long been one of the poorer provinces in China, lacking port access (river or maritime), and peppered with remote villages of various minority groups (such as the Miao, Buyi, Tai, and others). The province has also relied heavily on the state sector, especially in energy and mining. In the 1960s, Mao began a policy known as san xian “third front” which involved relocation of critical defense-related industries to the interior. Guizhou, along with Sichuan and Shaanxi was a major focus of the plan, making Guiyang into an industrial center. The city of Liupanshui 六盘水 in Guizhou West of Guiyang also became a major steel center from the plan. By the 2000s, inland provinces lagged far behind coastal China in economic growth during the reform and opening up period. Today inland China is seeing more growth and investment as China moves toward its “dual circulation” economy 国内国际双循环 and high prices in coastal cities drive youth to seek opportunities in second or third-tier cities.

Guiyang has not received much attention from scholars of urban China. But the city, built around numerous hills, feels like a sort of mini Chongqing—albeit without any buildings bisected by monorails. There are also a host of strange quirky buildings from the boom years of investment—a giant French chateau style mansion (called the White House) built against a mountain in central Guiyang1, a mall with a roof like a dolphin (and referred to as 海豚广场 dolphin plaza), and a “twin towers” development. Huaguoyuan 花果园, an extremely high-density residential urban renewal project that features tall towers wedged between mountains and large boulevards at the Western edge of the old area of Guiyang, is apparently China’s largest residential community with a population of around 400,000. As my host who spent years away from China but grew up in Guiyang told me, “all of this development has happened within a period of about 10 years, so its been jarring to return to the city, its totally different.”

Gui’an New Area

Today, the relocation of cloud computing to Gui’an New Area in Guizhou can be seen as the contemporary iteration of the third front. At a time of tension and rivalry between the West and China, idea of relocating critical digital infrastructure to the country’s interior made perfect sense. Guizhou, like other areas of the East-West Compute Plan, I wrote about here, has advantages in climate and energy resources that are also favorable for data center construction. At a higher elevation, Guizhou is cooler than much of the Southwest and there are ample hydropower resources.

When Gui’an New Area was approved as a National New Area by the State Council in 2014, plans to build it into a base for cloud computing were already underway. As early as 2012, provincial leader Li Zhanshu (a close former associate of Xi from Zhejiang) was pushing the province to seize national initiatives on data centers as a growth engine. Chen Min’er, another Xi associate, would become Guizhou’s party secretary. And Chen Gang, a former executive of Beijing’s Zhongguancun high tech cluster, would become Mayor of Guiyang. As described in this article by Kevin Liu of Guizhou University (2024) “the central and local government’s supportive policies, as well as the decisions made by key political figures and their networks, all played critical roles in Guizhou’s efforts to become a data centre cluster.” 2 Gui’an eventually attracted investment from Huawei (founder Ren Zhenfei hails from Guiyang originally), Tencent, China’s 3 telecom operators, Apple’s first China data center, and a Foxconn Factory.

In the last few years, the area has undergone some changes. The Foxconn Plant, an early anchor and provider of substantial employment, has for now shuttered. Data center construction continues, with new enterprise data centers of Midea, China Southern Power Network, China State Power Investment Corporation Limited. Data centers of China Construction Bank, and Bank of Communications are also under construction. However, the sprawling 1700 sq km- Gui’an New Area lacks the feeling of a vibrant urban area and is still mostly composed of relatively dispersed industrial parks, data centers, and a few isolated pockets of commercial and residential development. Villages that were not demolished exist in the interstitial spaces between newly built wide roads and highways, as written about by geographer Tim Oakes.3

A “seed bank” of Chinese civilization: The National Archives of Publications and Culture 中华文明的种子库： 国家版本馆

Earlier, I wrote about a key cultural project begun under Xi that has not gained much attention—either in China or abroad. This post is a follow up after I had a chance to visit two/four of the branches this past summer.

In 2019, Tsinghua University’s Zhuang Weimin (the Dean of the School of Architecture) received an invitation directly from the central publicity department of the CCP (中宣部) to participate in a competition to design the main branch of the new China National Archives of Publications and Culture 国家版本馆. The main branch is located in the mountains north of Beijing, directly on the city’s central axis. Unlike museums, which are often located in prominent locations and designed to attract the public, this is a different type of institution. It is remote, difficult to access, and focused primarily on archival and preservation work as well as serving as a “patriotic education base” for Party cadres and SOE employees. There are three other branches around the country: Hangzhou, Guangzhou, and Xi’an (with Beijing, each in one of the cardinal directions in the country). The system is described as a 1+3 cultural security system 文化安全体系, designed to safeguard China’s precious archives and artifacts. It is thus positioned as a critical part for Xi’s broader project of national rejuvenation and “cultural self confidence” 文化自信.

This past summer (2024) I visited two of the locations: Hangzhou and Xi’an. Only the Hangzhou location is currently open to the general public. The others are currently reserved for tours for 事业单位, essentially meaning govt or state-affiliated institutions. I was only able to visit the front of the Xi’an complex, which is located at the foot of the Qinling Mountains southwest of Xi’an.

While the project has received little attention either inside or outside of China, I’m particularly interested in it as a reflection of official thinking about the importance of cultural heritage and “cultural self confidence” to China’s current governance model, and to the nascent idea of the “two integrations” 两个结合—a Party propaganda effort to claim Confucianism and Marxism are perfectly compatible, and always were.1 How does the archive project do this? In some cases, quite literally. For example, in the main exhibit of the Hangzhou branch, a copy of the Yongle Dictionary 永乐大典 (completed originally in 1408 during the early Ming dynasty) is displayed alongside another exhibit on the history of the Communist Manifesto. In the Beijing branch, the main exhibit in the Wenhan Pagoda 文瀚哥 is “The Light of Truth - Classic Archives Exhibition of Marxism Localization and Modernization in China.” In the same room are copies of the Siku Quanshu 四库全书 (the Four Treasuries), Yongle Dictionary, and The Complete Classics Collection of Ancient China (completed in the Qing period under emperors Kangxi and Yongzheng). All of these encyclopedic editions were themselves scholarly and political projects of editing and curating history overseen by particular emperors, often with the goal of sanitizing history or erasing unwanted details from official records—in the case of Qianlong’s Four Treasuries, early conflicts between Han and Manchu that were seen as sensitive to the ruling Manchu class.2 The new National Archives of Publications and Culture both holds copies of these very editions, and serves as Xi Jinping’s own attempt to synthesize and curate Chinese history in a way that serves contemporary political ideology of the CCP.

In the past, these imperial “publications” were not seen as part of a continuous “canon” of Chinese literature, but this is indeed the implication of the exhibition and curation strategy of the NAPC.

“In the main exhibit of the Hangzhou branch, a copy of the Yongle Dictionary 永乐大典 (completed originally in 1408 during the early Ming dynasty) is displayed alongside another exhibit on the history of the Communist Manifesto.”

Hangzhou Branch 杭州分馆

The Hangzhou Branch, designed by Pritzker Prize winning Hangzhou-based architect Wang Shu 王澍 a stunning ensemble of architecture and landscape. It’s the only branch of the four currently open to the general public. Located in the suburban Liangzhu area of Hangzhou, the Hangzhou branch emphasizes Jiangnan style architecture and bibliographic culture. The Jiangnan Region, China’s most prosperous area in late imperial times and the crucible of literati culture, is a natural place to have a museum on China’s bibliographic and literary culture.

The main exhibit of the collection focuses on the bibliographic history of China and the region, and the numerous editions of important works: histories, maps, archives, and other documents. Interestingly, a copy of the Yongle Dictionary 永乐大典 is a highlight of the exhibit, and is placed in close proximity to an exhibit on the evolution of the Communist Manifesto. The implications are obvious-that the Communist Manifesto should not be viewed as in a category separate from important documents of China’s imperial past, they are both “editions” 版本 that reflect important moments in China’s history.

Xi’an Branch 西安分馆

The Xi’an Branch was designed by well known Xi’an architect Zhang Jinqiu 张锦秋, who has designed numerous cultural buildings in the city such as the Sha’anxi Provincial Museum 陕西省博物馆 and the recently completed Qin-Han Pavillion of the Sha’anxi Provincial Museum 陕西省博物馆秦汉馆. The location of the Xi’an branch is also interesting—note in the above image the central axis of the archive is aligned with the mountain peak behind it—a so-called “dragon’s vein” or longmai, in the traditions of Feng Shui. The area is also perhaps not coincidentally where illegal villas were demolished in a 2016 scandal that brought down much of Xi’an and Sha’anxi province’s top leadership. Xi Jinping, when visiting the area remarked, “let this be a warning that cadres must always protect the Qinling mountains.” The location of the archives built nearby these demolished villas, an investigation personally overseen by Xi, is yet another reminder of the power of the central state over the commercial interests of local officials.

Since I was only able to visit up to the front gate of the Xi’an branch (visits are currently limited to official state organs or 事业单位), I must rely on other news and reports of the site for insights into its content. The Xi’an branch is focused on important materials relating to the rich history of the area as the site of China’s earliest capitals (Chang’an in the Qin, Han, and Tang Dynasties) as well as the role of Xi’an as a gateway to the silk roads. I did speak to a friend who had visited, someone who works in the cultural industries in Xi’an, and he said the curatorial strategy was a bit sparse. “The project came out of nowhere, from the very highest reaches of the Party all of a sudden. But there hasn’t been a lot of thought into what it should be—a museum, a place for preservation? It’s still being worked out.” This might explain the current limited access to the Xi’an and the other two branches.

Conclusions

The ideological aims of the National Archives of Publications and Culture are fairly straightforward to anyone with even a cursory understanding of Chinese Marxist ideology, and the history of China’s imperial archival traditions. The project’s own reference to these earlier projects makes the connection even more clear. The NAPC obviously doesn’t include sensitive aspects of China’s recent history, but that is not unique—most state-backed institutions in China do not. Rather than censorship, the primary goal of the NAPC is to bolster the idea that the Communist Party is the rightful and legitimate inheritor of China’s ancient civilization, and to bridge the gap between the Party’s own ideas and history and the history of China’s imperial traditions—particularly in the world of publications.

Whether or how effective this project (and other related efforts) will actually be in rewriting China’s history remains to be seen. As Ian Johnson has written of China’s “underground historians”, there are many grassroots efforts to keep aspects of China’s history alive beyond the reach of the Party State—in the form of small groups of dissident archivists or digital archives.3 Can a project like the NAPC really transform the understanding of history with a vibrant, albeit heavily suppressed, online sphere?

As with much of China’s existing censorship and information control apparatus, the State knows such efforts will not reach everyone. But they don’t need to. By the financial backing of the state, institutions like the NAPC become prominent cultural facilities. Perhaps they will not usher in ideological transformation by themselves. But they do help in small ways to re-fashion the way the Party presents the past (both the distant past and its own more immediate history) to the public.

Because of their limited accessibility to the general public, the current audience of the project also should be seen in a more limited scope—the Party’s internal ranks of cadres, SOE managers, and officials. Friends of mine who have parents working as government officials had heard about their visits to various branches not open to the general public. Given Xi’s concern about the Party losing its ideological discipline and warnings to not “forget the original intention” (不忘初心), the NAPC project’s first audience seems to be the Party’s own rank and file. But whether Party discipline or ideological commitment can really be bolstered by such a complex seems a bit of a lofty aspiration in an age of widely available media and alternative ideas.