Two parallel AI infrastructure battles are escalating in Washington and Beijing, with potential knock-on effects for technology markets and energy demand. Republican Senator Josh Hawley called for a nationwide data center moratorium on November 12, 2026, while Reuters reported the next day that Chinese developers are planning more than $3.2 billion in AI projects built around restricted Nvidia chips.
Hawley, a Missouri Republican, announced the proposal on X as one of five priorities for Senate Republicans in the final six weeks before Democrats take control of the chamber. He wants a federal pause until Congress passes a “bill of rights giving taxpayers guarantees of no rate increases,” though he has not introduced detailed legislation specifying which projects would stop or how long the freeze would last.
The call comes after the Senate voted 73-25 on Monday to advance debate on the Bipartisan American Affordability and Jobs Act. That bill would fast-track federal energy infrastructure permits and require data center operators to cover the additional grid costs they create, rather than passing them to households. Proponents argue it would keep construction moving while protecting electricity customers; Hawley contends new construction should be halted until those protections are fully in place.
Electricity demand is at the center of the U.S. dispute. U.S. data centers consumed an estimated 183 terawatt-hours of electricity in 2024, and that figure could rise to 426 terawatt-hours by 2030, according to Pew data cited by Decrypt. Hawley previously introduced the GRID Act with Democratic Senator Richard Blumenthal in February, aiming to require new data centers to source their own electricity rather than relying on existing grid supplies. Independent Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez also support a national pause with broader rules covering AI’s effects on workers, public safety and the environment.
Public opposition is growing across party lines. A Pew Research Center analysis found 60% of Americans uncomfortable with a new data center in their area, including 54% of Republicans and 68% of Democrats. Some cities have already imposed restrictions. Meanwhile, BlackRock CEO Larry Fink warned last month that a construction moratorium could trigger a recession because so much investment now depends on AI infrastructure. OpenAI has promised to cover electricity and grid costs at future campuses it owns and operates.
In China, a Reuters investigation published Friday found more than $3.2 billion in AI data center projects specifying Nvidia H100, H200 or H800 processors, which are subject to U.S. export controls. The figure covers total project investment—including buildings, power and other infrastructure—not confirmed chip purchases or shipments.
Among the projects identified: a proposed 6.2 billion yuan ($873 million) computing center in Fujian province listing H100 servers but still seeking financing partners; a 3 billion yuan center in Shaanxi involving H100 and H200 chips; a 2.26 billion yuan project in Jilin combining Nvidia hardware with Chinese processors; and an Inner Mongolia proposal specifying 625 H200 servers. The documents do not establish that all the processors were obtained or that export laws were broken.
Washington began restricting advanced AI chip sales to China in 2022. In January 2026, the U.S. Commerce Department moved H200 export applications to case-by-case review, while H100 shipments remain more tightly restricted. That leaves Chinese developers choosing between familiar Nvidia systems that may be difficult to acquire and alternative processors that require software changes and additional engineering.
Nvidia still held 34% of China’s AI accelerator shipments in the first half of 2026, compared with Huawei’s 28%, according to an IDC estimate cited by Reuters. Chinese AI developer Z.AI has released an image-generation model trained entirely on Huawei chips, but the new project records show how deeply Nvidia remains embedded in Chinese AI infrastructure plans.
