In a Nov. 13, 2026, report, Reuters said it identified more than 23 billion yuan, or about $3.2 billion, in Chinese computing projects designed to use restricted US NVIDIA chips. The initiatives could face delays, supply disruptions or a switch to Chinese alternatives following NVIDIA co-founder Jensen Huang’s resignation as CEO and tighter oversight from Washington.

Reuters based its estimate on more than 150 procurement documents, investment plans and construction contracts. The projects range from large regional computing centers to retrofits of commercial facilities and involve NVIDIA accelerators including the H100, H200 and H800.

The $3.2 billion figure does not represent the cost of the chips alone. It includes planned investments, signed contracts and completed projects, as well as spending on buildings, power supplies and other infrastructure. Reuters could not determine exactly how all the processors mentioned in the documents entered China.

Chinese demand persists under export restrictions

The documents indicate that NVIDIA technology remains part of China’s AI infrastructure among universities, regional computing centers and projects involving local authorities, as well as major technology companies.

One project in Gansu province in northwestern China called for thousands of NVIDIA chips to be installed by the end of the year. It remains unclear whether the equipment was delivered.

Washington has restricted exports of advanced AI processors to China since 2022 and later extended the controls to modified chips, including the H800. In January, US President Donald Trump’s administration introduced case-by-case license reviews for H200 shipments to certain Chinese buyers, subject to specified conditions.

Estimates of NVIDIA’s share of the Chinese market vary by methodology. Huang said last year that the company’s share had fallen from 95% to zero. In an August report, NVIDIA effectively excluded revenue from data centers in China from its forecasts because of regulatory uncertainty.

An IDC estimate cited by Reuters in October found that NVIDIA accounted for 34% of AI accelerator shipments in China during the first half of 2026. The figures measure different parts of the market and do not establish that all the chips cited in the documents were imported legally.

Huang’s resignation adds uncertainty

Reuters published its findings one day after Huang stepped down as NVIDIA’s CEO amid a scandal involving his ties to Chinese businesswoman Alice Huang, whom US authorities suspect of possible links to Chinese intelligence. US lawmakers have called for tighter export controls on advanced chips.

Huang had argued for easing restrictions and emphasized the importance of the Chinese market to NVIDIA. His departure could weaken the company’s lobbying influence in Washington as lawmakers press for further controls.

Reuters separately highlighted NVIDIA’s Singapore-based cloud partner Megaspeed, which is linked to Alice Huang. Megaspeed denied illegally rerouting chips to China.

NVIDIA’s board has opened an independent investigation and said it had no reason to believe that corporate secrets or US national security had been put at risk. There have been no public findings that Huang violated export laws.

Reuters said the Chinese projects identified in its document review were not connected to Megaspeed or the allegations involving Huang.

Jamestown Foundation analyst Sunny Cheng said US regulators should examine not only the formal buyers of equipment but also the end users of the resulting computing capacity.

“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.”

US senators Elizabeth Warren and Jim Banks have also urged Congress to tighten restrictions on exports of NVIDIA’s advanced chips to China.

Source: Incrypted