Key Notes
- Reuters identified more than $3 billion in Chinese AI projects and equipment agreements designed around NVIDIA processors.
- The total includes planned investment, contracts and installations, covering infrastructure as well as hardware rather than chip sales alone.
- Huang’s departure adds uncertainty as lawmakers consider tighter export and remote-access controls, while NVIDIA promises continuity under interim CEO Colette Kress.
Jensen Huang’s departure from NVIDIA and mounting scrutiny in Washington are adding uncertainty to more than $3 billion in Chinese AI computing projects designed around the company’s processors, according to a Reuters investigation.
The news agency identified projects and equipment agreements worth more than 23 billion yuan, about $3.2 billion, after reviewing over 150 procurement notices, investment plans, construction contracts and corporate disclosures.
The total includes proposed investments, signed contracts and completed installations, as well as buildings, power systems and other infrastructure. It is not an estimate of NVIDIA chip sales, and it does not mean every project has secured the hardware it needs.
Demand survives the restrictions
Reuters found projects designed around H100, H200 and H800 processors across China, including local government-backed ventures, universities and regional computing centers. In one Gansu province project, thousands of NVIDIA chips were expected before year-end, but the report could not establish whether they had arrived.
That leaves operators exposed to delayed equipment, disrupted supply or the expense of adapting facilities for domestic alternatives. A plan to build AI capacity is only useful once the processors, servers and supporting systems can be brought together.
The H200 illustrates why buyers still seek these systems. NVIDIA lists 141GB of high-bandwidth memory and 4.8 terabytes per second of memory bandwidth for the accelerator, which is designed for AI training, inference and high-performance computing. Its server platforms combine multiple GPUs through high-speed interconnects.
Changing the processor can therefore mean more than replacing a component on an order form. Operators must consider software compatibility, networking and how their models run on the alternative hardware. Those decisions can affect the timetable and economics of a project already under construction.
NVIDIA promises continuity under Kress
NVIDIA announced Huang’s resignation as president, chief executive and a director on November 12. Chief financial officer Colette Kress is serving as interim CEO while the board considers internal and external candidates for the permanent role.
The company says its operations, customer commitments and product roadmap should continue without interruption. AIstify’s earlier coverage examined the transition and the infrastructure commitments Kress inherits.
Huang had been a prominent advocate for easing restrictions on sales to China. Reuters said his departure could weaken NVIDIA’s lobbying leverage as lawmakers press for tighter controls.
For Chinese customers, however, continuity inside NVIDIA cannot settle the separate question of access. Export licenses, enforcement and rules governing overseas computing services can determine whether a technically viable project is able to proceed.
Washington is looking beyond shipments
Reuters separately outlined proposals gaining attention in Congress as lawmakers push for tighter controls. The AI Overwatch Act would expand congressional oversight of export-rule changes, while the Chip Security Act would require location tracking and reporting of suspected diversion for advanced processors.
Another proposal, the Remote Access Security Act, would extend the Commerce Department’s authority over certain uses of controlled technology through overseas networks and cloud services. Reuters said the measure had passed the House in January and was awaiting Senate consideration. The proposals should not be confused with new restrictions already in force.
The distinction matters because a Chinese AI developer can rent GPU time abroad without importing a chip. AIstify reported on November 12 that firms were seeking to reserve overseas NVIDIA capacity in anticipation of possible tighter controls. That approach changes where the equipment sits; it does not remove the policy risk around who can use it.
The projects are separate from the allegations
Reuters said the Chinese projects it identified had not been linked to Megaspeed, the Singapore-based NVIDIA cloud partner, or to the allegations surrounding Huang. Megaspeed has denied illegally diverting NVIDIA chips to China.
NVIDIA’s board opened an independent inquiry on November 8 following reports that Huang had been targeted by a foreign intelligence effort. At the time, directors said they had no reason to believe company secrets or national security had been compromised.
The immediate issue for the projects is whether planned computing capacity can become usable infrastructure. Hardware demand remains visible in procurement records. Reliable access to that hardware is the less certain part.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
