Reuters Investigation: Jensen Huang's Departure and Export Control Review Deal Blow to Over 23 Billion Yuan in Chinese AI Computing Projects
A Reuters investigation of more than 150 public documents found that computing centers, university and enterprise data center projects based on Nvidia H100, H200 and H800 chips in coastal and western regions of China have a total value exceeding 23 billion yuan; this figure includes investment in data center buildings, power and supporting facilities, and is not entirely spent on purchasing chips. Jensen Huang's announcement that he is resigning as Nvidia CEO, combined with the U.S. export control review, may expose these projects to risks of supply disruption and delay. The report also noted that some Chinese companies are considering locking in Nvidia computing power at overseas data centers in advance, while the Remote Access Security Act (RASA) has passed the House of Representatives and, if enacted, would bring remote use of overseas computing power under regulatory scope.
Why read this
AI score 71
A Reuters investigation examined Chinese AI computing projects' reliance on Nvidia chips and the possible path by which export controls could extend to overseas computing rentals.
Topics
Body · AI translation
The investigation reviewed more than 150 public documents and found that related projects are spread across China's coastal technologically developed regions as well as western provinces, involving local government-supported computing centers, university and enterprise data centers. The Nvidia processors used or planned for use in the projects include models such as the H100, H200 and H800.
Reuters stressed that the more than 23 billion yuan is the total value of the related projects and equipment agreements, and does not mean that China purchased Nvidia chips of an equivalent amount. This figure also includes investment in data center buildings, power systems, supporting facilities and other infrastructure, while covering contracts already signed, projects under construction and investments still in the planning stage.
These projects reflect that Nvidia technology still has considerable influence in China's AI infrastructure construction. Even as the United States continues to tighten export restrictions on advanced AI chips, Chinese local governments, enterprises and research institutions are still advancing computing power construction based on Nvidia processors.
One noteworthy case is located in Gansu Province in northwestern China. A data center renovation project originally planned to install thousands of Nvidia chips before the end of the year, but it is currently unclear whether these processors have actually arrived. Some other projects have already begun operation or are under construction, while others may face delays due to chip supply problems.
As the United States further tightens export controls, the operators of these projects may have to face the risk of supply interruptions, construction delays, or even redesigning the entire computing infrastructure. If data centers originally designed around Nvidia H100, H200 or H800 architectures cannot obtain the corresponding chips, switching to domestic processors may require readjusting servers, software adaptation and cluster architecture, bringing additional costs.
Huang Renxun's departure may further increase this uncertainty. Reuters believes that Huang Renxun has long advocated relaxing chip export restrictions to China and is also an important policy communicator for Nvidia's business in China. His departure may weaken Nvidia's lobbying influence in Washington, giving U.S. lawmakers more room to push for stricter export controls.
The gap between U.S. export controls and China's actual demand
The United States began restricting exports of Nvidia's advanced AI processors to China in 2022, and subsequently tightened restrictions on chips specifically adjusted for the Chinese market, including the H800. In January 2026, the Trump administration began case-by-case review of applications from eligible Chinese customers to purchase H200 chips, with related conditions attached.
Nvidia previously stated in company filings that export controls have effectively forced it out of China's data center computing market. Huang Renxun also previously said that Nvidia's market share in China had fallen from 95% to zero.
However, Reuters cited data released by IDC in October this year pointing out that Nvidia still accounted for 34% of China's AI accelerator shipments in the first half of 2026.
These two figures are not directly contradictory because they measure different indicators: the market share Nvidia refers to concerns the company's business situation in China's data center market, while IDC's data reflects actual AI accelerator shipments. The latter cannot prove through what channels the restricted chips involved in the Reuters investigation actually entered China.
Reuters pointed out that it is currently impossible to determine whether the Nvidia chips required for these projects were obtained through legal approval, prior inventory, third-party supply channels or other means. The existence of the related projects also cannot directly prove that violations of U.S. export controls are involved.
The Huang Renxun incident draws further attention in Washington to chip flows
The investigation was published as Nvidia founder Huang Renxun had just announced his resignation as chief executive officer. Previously, his personal relationship with Alice Huang, a former executive of Singapore cloud computing company Megaspeed, was exposed, and the United States suspects that the latter has ties to Chinese intelligence agencies.
Megaspeed itself is also under U.S. scrutiny due to suspicions that it may have helped Chinese customers obtain Nvidia advanced chips subject to export controls. The company denies illegally transshipping chips to China.
Nvidia's board had previously launched an independent investigation and said there is no reason to believe that the company's trade secrets or U.S. national security interests have been harmed. As of the publication of the Reuters report, no public investigation conclusion had found that Huang Renxun violated U.S. export control laws.
In particular, it is necessary to distinguish that the Chinese AI computing power projects identified by Reuters this time have no evidence showing connections to Megaspeed or the allegations involving Huang Renxun. At present, the two are only simultaneously in the context of U.S. chip policy toward China and national security review.
Because the United States restricts the direct export of advanced chips to China, Chinese companies can still remotely use Nvidia processors deployed outside China by renting computing capacity from overseas data centers. This method allows companies to obtain the computing resources needed to train and run large AI models without directly importing chips.
Several people familiar with the matter said that a large Chinese cloud computing service provider has internally discussed whether to lock in overseas Nvidia computing power in advance, and some companies have also begun contacting data center operators in the hope of reserving related resources. Smaller AI labs are considering seeking cooperation with computing power leasing service providers in Southeast Asia.
These companies worry that the controversy surrounding Huang Renxun may push the United States to further tighten restrictions on China's use of overseas Nvidia computing power.
One focus is the Remote Access Security Act (RASA). The bill aims to give the U.S. Department of Commerce the power to regulate remote cloud access to restricted technologies. The bill has passed the U.S. House of Representatives but still awaits Senate review.
If the bill eventually becomes law, the way Chinese companies rent Nvidia chips through overseas data centers may also be subject to stricter regulation. This means that even if the chips do not directly enter China, as long as Chinese users can remotely use the related computing resources, they may also become subject to U.S. export control policy.
Overall, the Reuters investigation reveals a noteworthy reality: China's AI infrastructure construction still depends to a considerable extent on Nvidia technology, and the impact of U.S. export controls is no longer limited to whether chips can be directly exported, but may also extend to overseas computing power leasing, data center operations and remote access.
For China's AI computing power projects that have already invested huge amounts of money in construction, the real risk is not only whether chips can be bought, but also whether supply, maintenance and technical support can be continuously obtained in the future. If the United States further expands the scope of regulation, some projects may have to adjust construction plans and turn to domestic processors or other alternatives. However, there is currently no evidence that all the $3.2 billion projects identified by Reuters will be halted or canceled as a result.