Nvidia CEO Jensen Huang’s departure has raised concerns about the future of more than $3 billion in Chinese AI infrastructure projects. Several Chinese developers and research institutions have planned large computing facilities using Nvidia’s advanced H100, H200, and H800 processors.
A Reuters investigation published on November 13, 2026, identified Chinese AI projects and equipment agreements worth over 23 billion yuan ($3.2 billion). The investigation examined more than 150 procurement notices, investment plans, construction contracts, and corporate disclosures.
The report reveals China’s continued dependence on Nvidia chips despite tightening US export restrictions. Huang’s departure could add further uncertainty as Washington considers stricter controls on advanced AI processors and computing technology.
China’s $3.2 Billion AI Projects Depend on Nvidia Chips
A Reuters investigation identified several Chinese AI infrastructure projects across major technology hubs and less-developed western regions. These developments involve government-backed companies, universities, regional computing centres, and private technology developers.
The estimated $3.2 billion investment covers proposed projects, signed contracts, and completed installations. It also includes supporting infrastructure, such as data centre buildings and power systems, beyond Nvidia chip purchases.
Several projects specifically list Nvidia’s restricted H100 and H200 processors in their development plans. One proposed AI computing centre in Xianyou County, Fujian province, carries an estimated investment of 6.2 billion yuan ($873 million).
The project plans to use Nvidia H100 processors and is seeking commercial partners for development. Another computing centre in Shaanxi province involves a 3 billion yuan investment and specifies Nvidia H100 and H200 chips.
In northeastern China’s Jilin province, a separate 2.26 billion yuan project combines Nvidia H100 processors with domestically developed chips. Meanwhile, a proposed development in Inner Mongolia plans to install 625 Nvidia H200 servers.
These projects highlight the significant role Nvidia processors play in China’s planned artificial intelligence infrastructure. They also show how potential export restrictions could affect billions of dollars in planned AI investments.
Existing Chinese AI Data Centres Already Use Nvidia Chips
The report also uncovered Chinese AI computing projects that have progressed beyond the initial planning stage. Several facilities are already using Nvidia processors, while others remain under construction.
In Gansu province, a local company was awarded a 1.16 billion yuan construction contract in July. The planned computing centre is expected to house approximately 3,072 Nvidia processors, including H100 and H200 chips.
However, procurement documents do not confirm whether these processors have reached the facility. This uncertainty reflects the supply challenges facing Chinese developers under tighter US export restrictions.
Other Chinese computing facilities have already deployed Nvidia’s restricted processors. In March 2025, the Wuxi municipal government announced that an Nvidia H800 computing cluster had begun operations.
Shanghai-listed Damingcheng also disclosed an H800 computing cluster through one of its ventures in June. The company signed a computing services agreement initially valued at 759 million yuan. However, the agreement ended earlier than originally planned.
Chinese universities and research institutions have also purchased Nvidia’s advanced processors. Shenzhen University awarded a contract for three Nvidia H100 chips in September. Meanwhile, Northwestern Polytechnical University purchased two Nvidia H200 processors in October.
Jensen Huang’s Departure Creates New Challenges for Nvidia’s China Business
Jensen Huang’s departure comes as Nvidia faces growing scrutiny over its business connections and compliance with US export restrictions. The controversy has raised questions about the company’s relationships with China-linked partners and possible violations of American export controls.
The investigation involves Alice Huang, a businesswoman and former executive linked to Megaspeed. The Singapore-based company previously operated as a cloud partner for Nvidia. US authorities have reportedly examined her alleged intelligence ties to Beijing and possible chip diversion involving the company.
Megaspeed has denied illegally redirecting Nvidia processors to China. Nvidia’s board has launched an independent investigation into the matter. The company said it has no reason to believe its secrets or US national security were compromised.
However, no public findings establish that Jensen Huang violated US export-control laws. Reuters also found no links between the Chinese AI projects and the allegations involving Megaspeed.
Huang’s departure could also reduce Nvidia’s influence over technology policy discussions in Washington. He had repeatedly argued that American chipmakers should retain access to China instead of surrendering market share to domestic competitors.
His exit could make Nvidia’s efforts to influence future export policies more difficult. This could create additional uncertainty for the company’s access to China’s AI chip market.
ALSO READ: Nvidia CEO Jensen Huang Reportedly Targeted by Suspected Chinese Spy, US Intelligence Says
US Export Restrictions Add Pressure on Chinese AI Developers
The United States introduced restrictions on Nvidia’s advanced AI chips for China in 2022. Washington later expanded these rules to cover modified processors developed specifically for Chinese customers, including the H800.
In January 2026, the Trump administration introduced individual reviews for export licences covering Nvidia’s H200 processors. This approach could allow eligible Chinese customers to purchase these chips under specific conditions.
However, Nvidia has said US export restrictions have effectively pushed the company out of China’s AI data centre market. The company has also excluded Chinese data centre revenue from its financial forecasts because of ongoing regulatory uncertainty.
Despite these restrictions, Chinese developers continue to show strong demand for Nvidia processors. An IDC report cited by Reuters in October estimated Nvidia held 34% of China’s AI accelerator shipments during the first half of 2026.
This estimate differs from Jensen Huang’s earlier claim that Nvidia’s Chinese market share had fallen from 95% to zero. However, the figures measure different aspects of the market and do not explain how restricted processors reached Chinese customers.
Tighter US export controls could create further challenges for Chinese AI infrastructure projects. Developers may face construction delays, hardware shortages, and higher costs when replacing restricted Nvidia processors.
Companies could also need to redesign existing computing facilities around domestically produced chips. Such changes could increase development costs and delay China’s efforts to expand its AI computing capacity.
ALSO READ: Chinese AI Companies Rush to Secure Nvidia Chips as U.S. Restrictions Fears Grow
China’s AI Infrastructure Plans Face Increasing Uncertainty
China’s artificial intelligence industry faces growing challenges as developers expand computing infrastructure amid restrictions on foreign semiconductor technology.
Nvidia processors feature in several major Chinese AI projects, leaving developers exposed to changes in US export policies. Jensen Huang’s departure adds further uncertainty as Nvidia prepares to manage its China business under new leadership.
Meanwhile, China is investing in its domestic semiconductor industry to reduce reliance on American chipmakers. However, replacing Nvidia processors could require substantial investment, technical changes, and additional development time.
The $3.2 billion in identified AI projects does not mean every investment faces cancellation. Some computing facilities are already operational, while others remain under construction or await further development.
Their completion will depend on access to processors, changes in US export regulations, and alternative hardware availability. Chinese developers must also determine whether domestic chips can meet the computing requirements of these large-scale AI projects.
