Chinese artificial intelligence companies are trying to rent Nvidia-powered computing capacity at data centres outside China, Reuters reported on 12 November, citing three people familiar with the matter. The people said the companies are concerned that the governance turmoil at Nvidia and the scrutiny of Megaspeed could spur Washington to tighten restrictions on their remote use of advanced processors.
The bill they fear, the Remote Access Security Act, passed the House by 369 votes to 22 on 12 January 2026, according to the House Clerk’s roll call, and it awaits Senate action, Reuters reported. By our arithmetic, 304 days have passed since that vote.
Business Pill · GPUs Explained
An evergreen Business Engineer explainer on why AI runs on GPUs, the export-controlled chips at the centre of the remote-access debate. It teaches the general idea only and does not discuss this story.
The key insight: As we read it, the companies are moving on access rather than ownership: Reuters describes renting and reserving capacity in overseas data centres, the channel the Remote Access Security Act would bring under Commerce Department authority.
A major Chinese cloud provider has discussed internally whether to lock in overseas Nvidia chip capacity, while some companies have contacted data-centre operators about reserving it, the people told Reuters. They declined to identify the cloud provider because of the sensitivity of the matter.
Smaller AI labs are considering similar steps, including approaching major providers of rented computing power in Southeast Asia, two of the people said. The report, which we read in full in a syndicated copy, gives no volumes, prices or chip models for the capacity sought.
Reuters explained the gap at stake: Washington restricts exports to China of Nvidia’s most advanced AI chips, but Chinese companies can access such processors remotely by renting computing power at data centres overseas. RASA would give the Commerce Department authority to regulate remote, cloud-based access to restricted technology, Reuters wrote.
Nvidia, Megaspeed and the US Commerce Department did not immediately respond to Reuters’ requests for comment. Megaspeed has faced scrutiny over suspicions that it helped Chinese customers gain access to advanced Nvidia chips subject to US export controls, and the company has denied illegally diverting chips to China, Reuters reported.
Timeline of the Remote Access Security Act and related US export-control steps: the House passed H.R. 2683 by 369 to 22 on 12 January 2026 (Clerk roll call 13); BIS guidance of 31 May 2026; Nvidia’s 10-Q flagging the bill, filed 26 August 2026; the Reuters report of 12 November 2026; and the end of the affiliates-rule suspension on 10 January 2027 (91 FR 70789). Day counts are our arithmetic.
Grey-Market Demand
The prospect of tougher enforcement has also fuelled demand for restricted Nvidia AI chips in China’s grey market, prompting traders and intermediaries to start stockpiling these processors, two of the people said. Buyers fear a broader US push against smuggling of advanced chips and AI servers into China, they told Reuters.
Reuters also cited procurement records showing that a Chinese university with deep ties to the People’s Liberation Army bought Nvidia’s H200 chips last month despite being on a US export blacklist, a purchase set out in our piece on the H200 buyer. US prosecutors in March charged a Super Micro Computer co-founder and two others with conspiring to divert servers containing Nvidia chips to China through a Southeast Asia-based company, Reuters noted; the defendants have not been convicted.
The Bill and the Vote
The Clerk’s roll call 13 records H.R. 2683 passing on a motion to suspend the rules, with 369 yeas, 22 nays and 39 members not voting. Republicans voted 167 to 22 and Democrats 202 to 0. By our arithmetic, the yeas were 94.4% of the 391 members who voted yea or nay.
The House Select Committee on the CCP said that day the bill addresses the issue “by clarifying that export controls can be applied equally to remote access, provision of access, and cloud-based exposure of controlled items”. We reported on 10 November on the White House position on closing remote access.
Bureau of Industry and Security guidance of 31 May 2026 says a license is required to export advanced computing items to entities headquartered in Country Group D:5 or Macau “even if the entities themselves are located outside Country Group D:5 or Macau”. The same guidance says bona fide data-centre operators are not required to cease the ongoing use of such items because of it, “until further notice from BIS”.
Analyst view (12 Nov): one forecast of what comes next. Andrew Curran (@AndrewCurran_) quote-posted a summary of the Reuters report on X on 12 November, adding a prediction of the sequence: “This is what’s going to happen, they are going to close the remote access loophole next week. After that, they will pass bipartisan chip export controls to China. After that, Biden style global diffusion laws, only stricter. After that compute strategic resource, US only.” This is Curran’s prediction, not a reported fact: no vote on remote access has been scheduled in any primary record we have seen, and the bill described above still awaits Senate action.
What Nvidia Tells Investors
Nvidia’s 10-Q, filed on 26 August 2026, says: “Congress is also considering legislation such as the Remote Access Security Act, or RASA, which could prohibit the provision of cloud services to any company with an ultimate parent headquartered in China.” It adds that, if enacted, RASA “could have a material impact on our business, operating results, and financial condition”.
The same filing puts revenue from customers headquartered in China, including Hong Kong, at $7,880 million of $96,221 million in the quarter to 26 July 2026. By our arithmetic, that is 8.2% of the quarter’s revenue, against 8.5% a year earlier. Our search of SEC filings found the bill named by eight companies, including Nvidia and Alibaba, as set out in our piece on the cloud-access bill in filings.
Nvidia revenue by customer headquarters location in the quarter to 26 July 2026, from its 10-Q filed on 26 August 2026: United States $60,074 million, Taiwan $26,985 million, China including Hong Kong $7,880 million and other $1,282 million, of $96,221 million in total. Shares are our arithmetic.
Where the Truce Stands
The capacity rush comes as the 2025 trade truce runs on short extensions. BIS rule 91 FR 70789 extends the suspension of the affiliates rule through 10 January 2027, 59 days from 12 November by our arithmetic, as set out in our piece on the four truce terms ending in January. Axios reported on 11 November that Trump has agreed to allow sales of Nvidia’s H200 chips to China in exchange for a 25% tariff, as that piece notes.
As we read the dates, the bill passed the House 304 days ago by our arithmetic, and Nvidia has told investors since at least its 26 August 10-Q that, if enacted, it could have a material impact on the business.
As we read the filing, revenue from customers headquartered in China, including Hong Kong, was 8.2% of Nvidia’s revenue in the quarter to 26 July 2026 by our arithmetic; the table counts sales by customer headquarters, so it does not measure capacity rented abroad.
Rep. John Moolenaar (R-Mich.), Select Committee on China chairman, in the committee release of 12 January 2026
“This bill brings our laws into the digital age and makes it clear that cloud compute is subject to U.S. export control law, just like physical chips.”
Three Implications
FOR CHINESE AI DEVELOPERS Restricting overseas cloud access would make it harder for them to obtain high-end computing power without importing the chips directly, Reuters wrote.
FOR NVIDIA Its 10-Q says RASA could prohibit the provision of cloud services to any company with an ultimate parent headquartered in China.
FOR CLOUD OPERATORS BIS guidance says bona fide data-centre operators need not cease the ongoing use of advanced computing items because of it, until further notice from BIS.
The framework argues that the AI supply chain is “creating geopolitical leverage that operates independently of who has the best models”.
As we read it, rented capacity abroad is one place that leverage now sits: the chips stay in third countries, and the bill would let Washington regulate access to them. The framework predates this report and does not discuss it.
What Is Not Established
Reuters did not name the companies seeking capacity, and its sources were not identified. How much capacity was reserved, at what price and from which providers is not established. The Senate has not voted on RASA, Reuters reported, and the Commerce Department did not immediately respond to its request for comment.
The 10-Q figures classify revenue by customer headquarters location and cover sales, not rented capacity. We did not contact Nvidia, Megaspeed, the Commerce Department or any cloud provider for this piece.
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.
Chinese AI companies are trying to rent Nvidia-powered capacity at overseas data centres before any US move on remote access, Reuters reported, citing three people familiar with the matter. The bill that would bring cloud access under export controls passed the House 369 to 22 and, 304 days later by our arithmetic, awaits Senate action.
A note on sourcing. We read the Reuters exclusive of 12 November in full in a copy syndicated by MarketScreener; its sources are three people familiar with the matter whom Reuters did not identify. We read the House Clerk’s roll call record and the House Select Committee’s release on the Remote Access Security Act, the relevant sections of Nvidia’s 10-Q filed on 26 August 2026, the BIS guidance of 31 May 2026 in full and the Federal Register summary of 91 FR 70789.
The Axios report is as cited in our earlier coverage. The day counts and percentages are ours. Under our editorial policy, this piece names no private individuals and does not reproduce Reuters’ headline wording or its paragraph on personal matters. We did not contact Nvidia, Megaspeed, the Commerce Department or any cloud provider.
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.
