Huawei bears the heaviest burden; sources cite DeepSeek’s V5 training on Ascend

The Commerce Department said this week that it was closing loopholes that allowed Chinese companies to bypass export restrictions on Nvidia’s most powerful AI chips by remotely accessing them through data centers outside China, especially in Southeast Asia. The rules target companies such as Megaspeed, which built data centers in the region and arranged for billions of dollars’ worth of top-of-the-line Nvidia chips to be shipped to China, according to people familiar with the situation.

New rules pressure financing and model development

The Wall Street Journal reported that the new rules were issued in the aftermath of an affair between Nvidia founder Jensen Huang and former Megaspeed executive Alice Huang. The revelation of the affair invited an investigation by the Securities and Exchange Commission.

The rules are already affecting financing. Bankers are reassessing plans to finance Nvidia chip purchases by Singapore-based Aolani Cloud that would provide computing power to Chinese tech giant ByteDance, according to a person familiar with the matter. An Aolani spokesman said the company was working to ensure it remained compliant, while China’s Ministry of Commerce called the rules an abuse of export controls.

Computing shortages are already apparent in China. Since mid-November, DeepSeek has restricted its chatbot, including limits on the length of documents it will read, while reallocating computing resources to prioritize training workloads, according to people familiar with the matter. Moonshot AI suspended new subscriptions shortly after releasing its flagship Kimi K3 model in July, then apologized in October for limiting usage after releasing the Kimi K3.1 Pro update.

Engineers working for DeepSeek, Moonshot AI and Alibaba said the pace of research, especially the rollout of new models, has repeatedly slowed over the past year, partly because of computing-capacity shortages. DeepSeek and Moonshot AI did not respond to requests for comment, and Alibaba declined to comment.

Top Chinese labs remain reliant on Nvidia hardware

The Wall Street Journal reported that, despite a Chinese government push to promote domestic chips, China’s top AI labs have continued to rely on Nvidia’s most advanced chips to build their flagship models, especially during intensive early-stage training. U.S. restrictions on exports of those chips to China led many of China’s top AI labs to access them through data centers in places such as Malaysia and Indonesia to train their models. The workaround may no longer work, The Wall Street Journal said.

The Wall Street Journal reported that the move to cut China off from high-end American computing power would hurt Beijing’s AI prospects, particularly in the near term. The newspaper said China matches or beats the U.S. in many of the factors that drive progress in AI, including engineering talent and access to energy, but remains at least 10 years behind the U.S. in making the best chips. The restrictions also come as The Wall Street Journal reported that China’s AI industry is struggling to generate profits.

“The new regulations will close the door to most Chinese lab remote access,” said Daniel Nishball, director of research at SemiAnalysis, an AI-infrastructure data and consulting firm. Chinese AI companies are already exploring ways to access data centers in the U.S. and Europe through local entities without obvious ties to China, according to industry insiders. The insiders said such arrangements will be risky and that finding clusters of computing power big enough to train new models will be difficult.

China has built several large domestic computing clusters using smuggled hardware, including Nvidia’s Blackwell chips, but insiders said the clusters were far from enough to meet AI training demand.

Huawei’s domestic capacity remains far below demand

The Wall Street Journal reported that the task of filling the computing gap would fall most heavily on Huawei Technologies, the Chinese tech-equipment giant and producer of China’s most powerful domestic AI chips. Huawei has used networking expertise from its telecommunications business to bundle together hundreds of thousands of AI chips into clusters that Huawei says approach the performance of cutting-edge American silicon.

People familiar with the matter said DeepSeek has been training its next flagship model, V5, on Huawei’s Ascend chips. If V5 is successfully trained mainly with Huawei hardware, the people said, that would indicate Chinese domestic chips are capable of large-scale frontier AI training. The Information first reported DeepSeek’s plan to rely on Huawei hardware to train V5.

Huawei said in September that it plans to introduce two new AI chips next year, with further releases planned for 2028 and 2029. Specifications released by Huawei and Nvidia indicate that the computing power those chips can deliver will still trail Nvidia’s best products.

Even if the reported DeepSeek-Huawei collaboration proves successful and the new chips arrive on time, Huawei faces another problem. U.S. export controls on chip-making equipment mean China struggles to produce high-quality AI chips in large quantities. Huawei was blacklisted by the U.S. and faces even stricter limits on access to chip-making tools, and Huawei’s planned production of its latest chip — around 750,000 this year — falls far short of demand.

Huawei has installed many of those chips in a sprawling facility in Inner Mongolia’s Horinger County. The facility is one of the largest data centers in China, but Epoch AI’s recent analysis put its computing power at around 7% of SpaceX’s Colossus 2, the largest data center in the U.S., in Tennessee. Epoch AI is a U.S.-based research firm.

Lawmakers consider tighter chip export controls

Chinese access to U.S. chips could get even tighter. American lawmakers are considering new export controls on chip technology and chip-making equipment, including giving Congress the authority to block specific export licenses. On Wednesday, the congressional U.S.-China Economic and Security Review Commission recommended new rules blocking any export of chips to China that would divert chip manufacturing or AI capacity from the U.S.

Critics, including former Nvidia CEO Jensen Huang, have argued that China should remain reliant on American technology rather than developing its own chips outside U.S. control. Proponents of controls say Chinese leader Xi Jinping has prioritized developing domestic chip supplies as a matter of national security.

“The Chinese have decided the course of action here, which is they intend to either steal the technology or build the technology, or both,” said Mike Kuiken, vice chair of the U.S.-China Economic and Security Review Commission. “They cannot, in their view, be dependent on the democratic world’s technologies.”

Tightening access to American technology could accelerate China’s timeline by encouraging domestic tech companies to devote more serious attention to developing chips, said Rui Ma, founder of Tech Buzz China, a U.S.-based research firm. But given how far China remains behind the U.S. in chip technology, Ma said pressure from losing access to American computing power is unlikely to produce immediate progress. “It is going to take a while,” Ma said.