Nvidia has said US export controls effectively shut it out of China’s AI data-centre market. A Reuters review published on 13 November found Chinese computing projects and equipment agreements built around its restricted H100, H200 and H800 processors, with a combined reported value of more than 23 billion yuan ($3.2 billion). Reuters examined more than 150 procurement notices, investment plans, construction contracts and corporate disclosures.
The figure covers planned investments, signed contracts and completed installations, including buildings and power systems, and does not reflect spending on Nvidia chips alone, Reuters says. In a companion factbox, Reuters named five projects with a stated value. They add up to 14.12 billion yuan, about 61% of the 23 billion yuan total by our arithmetic, and the largest is a proposal still seeking partners.
Business Pill · Embedded Dependency
An evergreen Business Engineer explainer on embedded dependencies: what is inside what you buy. It teaches the general idea only and does not discuss the Reuters review.
The key insight: A Reuters review of more than 150 Chinese documents found more than 23 billion yuan ($3.2 billion) of AI computing projects built around restricted Nvidia H100, H200 and H800 chips; the five projects Reuters itemised with a value add up to 14.12 billion yuan, about 61% of the total by our arithmetic.
The projects run from China’s coastal technology hubs to remote western provinces, Reuters reported, and reach beyond the largest tech firms into local government-backed ventures, universities and regional computing centres. The documents show Chinese customers continuing to plan purchases and, in some cases, operating restricted processors, according to Reuters.
Reuters is careful about what its documents show. On the share figures it writes: “The figures measure different things and do not establish how the restricted hardware identified in the Reuters investigation reached Chinese users.” Reuters says China’s commerce ministry and Nvidia did not immediately respond to its request for comment.
Sunny Cheung, an analyst at the Washington-based Jamestown Foundation, told Reuters the findings reinforce the analyst’s own view on enforcement: “Washington must stop focusing solely on who buys the chips on the surface and start examining who ultimately uses the computing power—and for what purpose.”
The five Chinese AI computing projects Reuters itemised with a stated value on 13 November 2026, in billion yuan, by stage. Values include buildings and power, not chips alone. The 14.12 billion yuan total, the 61% share of 23 billion yuan and the stage split are our arithmetic.
Five Projects, 14.12 Billion Yuan
The largest is a proposed 6.2 billion yuan ($873 million) AI computing centre in Xianyou county, Fujian province, which seeks commercial partners and specifies Nvidia H100 chips, Reuters reported. In Shaanxi province, state-owned engineering group Sinomach said a subsidiary won a contract as part of a project for a 3 billion yuan computing centre specifying H100 and H200 processors.
In Jilin province, an April government prospectus describes a 2.26 billion yuan project combining H100 processors with Chinese-developed chips; the developer expects bank loans to finance about 80% of the investment. In Inner Mongolia, a 1.5 billion yuan development proposes installing 625 Nvidia H200 servers, and Chinese industry reporting in June said construction had begun on two buildings in its first phase.
In Gansu province, a local company awarded a 1.16 billion yuan construction contract in July for a computing centre that plans to install approximately 3,072 Nvidia processors, according to tender documents specifying H100 and H200 hardware, Reuters reported. The contract went to a consortium led by a subsidiary of state-owned Shaanxi Construction Engineering. Reuters adds: “The documents do not establish whether the processors have been delivered.”
By our arithmetic, 8.46 billion yuan of the 14.12 billion, about 60%, sits in the two projects at proposal or prospectus stage (Fujian and Jilin). The other 5.66 billion is in the three with a contract won or construction under way.
Clusters Already Running, Down to Two Chips
Other facilities are already using restricted Nvidia hardware, Reuters reported. In Jiangsu province, the Wuxi municipal government said in March 2025 that a computing cluster using Nvidia H800 processors had entered operation. In Fujian, Shanghai-listed Damingcheng disclosed in June that a computing venture had built an H800 cluster; its computing-services agreement was originally worth 759 million yuan, although the contract ended early.
The orders run small too. Shenzhen University awarded a contract for three H100 processors in September, and US-blacklisted Northwestern Polytechnical University bought two H200 chips the following month, Reuters reported separately; we covered that H200 purchase on 11 November.
Zero Share, or 34%?
Jensen Huang, a vocal advocate of easing export restrictions, said last year that Nvidia’s market share in China had fallen from 95% to zero, Reuters reported. Nvidia’s August filing said it was effectively excluded from China’s data-centre computing market, and the company has excluded China data-centre computing revenue from its forecasts, citing regulatory uncertainty.
An IDC report cited by Reuters in October estimated that Nvidia still accounted for 34% of Chinese AI-accelerator shipments in the first half of 2026. Reuters says Nvidia’s statements address its own ability to compete and sell directly in China, and that the two figures measure different things.
Controls, Lobbying and the Week’s Context
Washington began restricting Nvidia’s advanced AI processors in 2022 and later tightened controls on modified chips developed for China, including the H800. In January, the Trump administration moved to review H200 export licences for eligible Chinese customers individually, subject to conditions, Reuters reported.
The review was published a day after Huang resigned as chief executive. Reuters says the departure could weaken Nvidia’s lobbying leverage as US lawmakers press for tighter controls; OpenAI backed three chip export bills this week. Reuters also writes: “There has been no public finding that Jensen Huang violated export-control laws.” The Chinese projects it identified have not been linked to Megaspeed or to the allegations involving Huang, Reuters says. For the Singapore side, see what we know about Megaspeed and the fifth person charged in Singapore.
As we read the two Reuters reports together, the money is mostly in plans rather than racks. About 60% of the 14.12 billion yuan Reuters itemised, by our arithmetic, sits in a Fujian proposal still seeking partners and a Jilin prospectus that expects bank loans to fund about 80% of its investment. Those are the projects that, on Reuters’ account, face delays or costly redesigns around domestic chips if export controls tighten.
Reuters, 13 November 2026
“The figures measure different things and do not establish how the restricted hardware identified in the Reuters investigation reached Chinese users.”
Three Implications
FOR NVIDIA Nvidia’s August filing said it was effectively excluded from China’s data-centre computing market, while an IDC report cited by Reuters put it at 34% of Chinese AI-accelerator shipments in the first half of 2026.
FOR CHINESE PROJECT DEVELOPERS Reuters says operators are exposed to potential delays, supply disruptions or costly redesigns around domestic chips as export controls tighten.
FOR WASHINGTON Reuters quotes the Jamestown Foundation’s Sunny Cheung saying enforcement should examine who ultimately uses the computing power, not only who buys the chips.
The framework says: “This isn’t just a supply chain problem. It’s a fundamental constraint on the speed of AI transformation, and it’s creating geopolitical leverage that operates independently of who has the best models.”
As we read it, the Reuters review shows that leverage in procurement paperwork: tenders and prospectuses in at least five provinces written around one American supplier’s restricted chips. The framework predates the Reuters review and does not discuss these projects.
Nvidia share figures as Reuters reported them on 13 November 2026: Jensen Huang said last year the China share fell from 95% to zero; an IDC report cited by Reuters in October put Nvidia at 34% of Chinese AI-accelerator shipments in the first half of 2026. Reuters says the figures measure different things.
What Is Not Established
We did not see the procurement documents; every project figure comes through Reuters. The 23 billion yuan total mixes plans, contracts and installations and includes buildings and power, so it is not a count of chip spending. Reuters says its documents do not establish whether the Gansu processors were delivered, or how the restricted hardware reached Chinese users. Nothing in the reporting we read says any named project broke US rules.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
A Reuters review of more than 150 Chinese documents found more than 23 billion yuan ($3.2 billion) of AI computing projects built around restricted Nvidia chips, from a 6.2 billion yuan proposal in Fujian to two H200 chips at a university. The five projects Reuters itemised with a value come to 14.12 billion yuan, about 61% of the total by our arithmetic, and about 60% of that is in projects at proposal or prospectus stage.
A note on sourcing. We read Reuters’ 13 November investigation and its companion factbox in full through their MarketScreener syndications; we first saw the report relayed in a post on X. We did not see the Chinese procurement notices, contracts or disclosures; every project figure, the IDC estimate and the Jamestown Foundation quote come through Reuters, as does the note that Nvidia and China’s commerce ministry did not immediately respond to Reuters.
The 14.12 billion yuan total, the 61% share and the 60% stage split are our arithmetic on Reuters’ figures. We did not contact Nvidia or the companies and institutions named.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
