Chinese AI Firms Rush to Lock Up Overseas Nvidia Capacity
Chinese AI firms are reserving overseas Nvidia compute and stockpiling grey-market chips, fearing the Huang controversy will boost the Senate-pending Remote Access Security Act.

Why it matters
- Reuters reported on November 12, 2026 that Chinese AI firms are seeking rented overseas Nvidia capacity, citing three sources.
- The House passed the Remote Access Security Act 369-22 on January 12, 2026; the Senate received it January 13 and referred it to the Banking Committee.
- RASA would amend the Export Control Reform Act of 2018 to let Commerce regulate foreign persons' remote access to U.S.-jurisdiction items.
- A major Chinese cloud provider has internally discussed locking in overseas Nvidia capacity; smaller AI labs have approached Southeast Asian compute providers.
- Grey-market traders and intermediaries in China have begun stockpiling restricted Nvidia AI chips.
By Sophie Lindqvist
Chinese AI companies are trying to lock in rented Nvidia chip capacity at overseas data centers because they fear the controversy around Nvidia CEO Jensen Huang could trigger tougher U.S. limits on remote access, Reuters reported on November 12, 2026, citing three people familiar with the matter.
The report lands at a moment when Washington's chip rules already block exports to China of Nvidia's most advanced AI processors. But the rules govern the physical shipment of chips. They don't clearly cover a Chinese company renting time on those same chips inside a data center in Southeast Asia. That gap is what Chinese firms now worry will close.
What exactly are the Chinese companies doing?
According to the report, a major Chinese cloud provider has internally discussed locking in overseas Nvidia capacity. Some companies have gone further and approached data-center operators about reservations. The sources withheld the provider's name because of the matter's sensitivity.
Two of the three sources said smaller Chinese AI labs are weighing similar steps. Their approaches include talks with large rented-compute providers in Southeast Asia.
None of this involves importing chips. It's all contract work — reserving computing power that happens to run on Nvidia silicon sitting outside China's borders. The takeaway: Chinese firms are trying to secure advanced computing through contracts rather than hardware, before the rules catch up.
Why does remote rental sidestep export controls?
The mechanism is simple once you separate the chip from the service it provides. U.S. export rules restrict shipping a processor to China. They don't, in current practice, restrict a Chinese developer from connecting over the internet to that processor housed abroad.
Those processors train and run the large models behind chatbots and other generative-AI products. For Chinese developers, remote rental has become a practical supply route: high-end computing without a single chip crossing the border.
That's also why the route is fragile. If Washington restricts overseas cloud access, high-end computing becomes much harder for Chinese developers to obtain without importing chips directly — and direct imports of the most advanced Nvidia AI chips are already restricted. In plain terms: close the cloud loophole and the workaround disappears.
What is the Remote Access Security Act?
The companies' specific fear, according to the sources, is that the Huang controversy could add political momentum to the Remote Access Security Act, known as RASA. The bill would amend the Export Control Reform Act of 2018 — the core U.S. export-control statute — to cover remote access by foreign persons to items under U.S. jurisdiction. It would also authorize the Commerce Department to regulate that access.
RASA's definition of covered remote access is broad. It covers access through a network connection, including the internet or a cloud computing service, from a location other than the item's physical location.
The bill applies that definition to access that is purposeful, knowing, reckless or negligent by a foreign person. But there's a gate: the provisions kick in only when the Commerce secretary determines that use of the item could pose a serious risk to U.S. national security or foreign policy. The definition does not reduce the mental-state requirement for criminal liability.
Where does the bill stand?
RASA has passed one chamber but is not law. The House passed the bill on January 12, 2026, by a vote of 369-22. The Senate received it on January 13, 2026, and referred it to the Committee on Banking, Housing, and Urban Affairs. Congress's tracker still lists the measure as "Passed House."
So the proposed remote-access authority remains pending. The reservation activity Reuters describes is a bet on which way that vote goes. One line captures the legislative state of play: overwhelmingly popular in the House, parked in a Senate committee, and not yet binding on anyone.
What would RASA require of Commerce?
The bill also builds in a consultation requirement. It would require Commerce to keep the House Foreign Affairs Committee and the Senate Banking Committee informed about anticipated regulations.
That reporting would cover three things, per the bill: the security risk, the regulatory method, and possible effects on the U.S. economy. The consultation provision does not require either committee's approval before Commerce uses the proposed authority.
That last detail matters for companies planning around the bill. If RASA becomes law, Commerce could move to regulate remote use without waiting for a green light from Congress — informed, but not constrained, by the committees. In short: the oversight is informational, not a veto.
Are grey-market buyers also reacting?
Yes, and the reaction extends beyond cloud contracts. Two of the sources said the prospect of tougher enforcement has increased demand for restricted Nvidia AI chips in China's grey market. Traders and intermediaries have started stockpiling processors.
The buyers' reasoning mirrors the cloud providers'. They fear the Huang controversy could produce a broader U.S. crackdown on smuggling advanced chips and AI servers into China. Stockpiling is the hardware-side hedge against that outcome; overseas reservations are the contract-side hedge.
Both moves point to the same underlying calculation. Chinese AI developers read the political environment and concluded the window for relatively easy access to top-tier Nvidia computing — whether rented abroad or bought in the grey market — may be closing. The takeaway: precaution is now the operating mode on both fronts.
Why does this matter beyond China?
The story shows how export controls are colliding with the cloud-era shape of computing. When advanced AI runs on rented processors in distant data centers, the physical border stops being the natural checkpoint. RASA is Congress's attempt to move that checkpoint to the network connection.
For Chinese AI developers, overseas capacity is both a supply route and a policy exposure. The reported reservation efforts and the stockpiling are precautionary responses taken while RASA remains pending, not responses to a law already in force.
If the bill becomes law, Commerce would gain authority to regulate remote use of U.S.-jurisdiction items — making overseas compute a potential export-control issue even when the chips never enter China. Until the Senate acts, every reservation signed in Southeast Asia is a hedge against a vote that hasn't happened.
Source: unite.ai
More from Sophie Lindqvist
Show full bio
Staff writer covering regulation, policy and their impact on artificial intelligence and machine learning at AI In Context.
548 articles
Related articles
- Smuggled Chips and Server Loopholes: How China Gets Nvidia GPUs
- Moonshot AI Sought Nvidia Blackwell Capacity in Europe, Sources Say
- Jensen Huang says Nvidia must work with China on AI safety
- Blacklisted Chinese Defense University Bought Four Nvidia H200 Cards
- Report: Jensen Huang's Affair Linked to $7.5 Billion in Chips Reaching China