What the New Rule Actually Says
The U.S. Commerce Department's Bureau of Industry and Security (BIS) has issued an interim final rule that fundamentally changes how it controls access to high-powered AI chips. Previously, restrictions focused on the physical export of advanced semiconductors.
This new rule extends those controls to the service of using those chips. It requires data centers anywhere in the world to get a U.S. license if they provide significant computing power to companies headquartered in or owned by parent companies in China and other restricted nations. This effectively treats renting powerful AI computing in the cloud much like shipping the chips themselves.
Closing a Critical Loophole
The primary goal of this new regulation is to prevent the circumvention of existing export controls. While the U.S. had already blocked the direct sale of top-tier AI chips, like Nvidia's advanced models, to China, a significant gap remained. Chinese companies could still legally access the same computing power by renting it from cloud providers or data centers located in third-party countries. Lawmakers and policy experts argued this allowed restricted entities to 'access our tech through the cloud,' rendering physical export bans less effective. This new rule is designed to plug that exact hole by focusing not just on where a chip is sent, but who is ultimately using its power.
The Impact on Big Tech and Global Clients
Major U.S. cloud providers like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud are now at the forefront of enforcing this policy. They, along with other data center operators worldwide, face significant new compliance obligations. They must now implement 'Know Your Customer' (KYC) programs to verify the identity and ownership of clients training large AI models. The rule also creates new licensing requirements for shipping advanced chips to several countries identified as having a 'higher-than-normal risk of diversion,' including Singapore, Malaysia, Thailand, Indonesia, and Brazil. This puts a new burden on data centers in these booming tech hubs, potentially affecting their appeal as neutral grounds for AI development.
A New Chapter in the Tech Cold War
This move is a clear escalation in the broader geopolitical and technological rivalry between the U.S. and China. The stated aim is to undermine the ability of companies in 'countries of concern' to train advanced AI models that could be used in ways that threaten U.S. national security. By shifting the focus from hardware to computing services, the U.S. is treating access to compute as a controlled commodity. While the U.S. sees this as a logical extension of its national security strategy, Beijing has accused Washington of arbitrarily suppressing the development of China's high-tech industries. This action will likely accelerate China's efforts to develop its own domestic chip and cloud infrastructure, further fragmenting the global tech ecosystem.
Challenges and Unanswered Questions
Enforcement of this new rule presents considerable challenges. It requires data center operators to monitor not just their direct clients but also the ownership structures behind them to ensure they aren't controlled by entities from restricted nations. There are questions about how effectively cloud providers can track how their infrastructure is being used without violating customer privacy. The rule does include exemptions for commercial recommender systems used by online retailers and social media, which are not seen as posing the same national security risk. However, for data centers that are already serving restricted companies, they have a 90-day window to obtain the necessary licenses to continue providing those services.
















