Major Cloud Providers Take on New Role in AI Chip Access

The United States is set to empower major tech companies like Google and Microsoft to serve as gatekeepers for access to highly sought-after AI chips around the globe. This move is part of a broader strategy to limit Chinese access to advanced technology.

The anticipated new regulations, which could be announced as soon as this month, will require these companies to meet strict criteria. They will have to report critical information back to the US government and prevent Chinese entities from obtaining AI chips. This regulation will enable approved firms to provide artificial intelligence services over cloud platforms internationally without needing a special license.

These developments highlight the urgency among US officials, especially as the Biden administration approaches its final days. They aim to expedite the approval process for exporting AI chips while simultaneously restricting access to potentially harmful actors. The US government is concerned that China might leverage AI technology for military advancements, significant cyberattacks, or the development of bioweapons.

While the US Commerce Department has not commented on the specifics or timing of these new regulations, sources indicate that the plans may still be subject to change. Both Google and Microsoft have yet to respond to inquiries about the forthcoming measures.

This approach follows a national security agreement Microsoft signed with the US government in April, which allows them to provide AI technology to the Emirati company G42. According to the new draft regulations, other companies that do not have gatekeeper status will still be able to apply for licenses to import a limited number of high-end AI chips from companies like Nvidia and AMD, depending on the country.

Nvidia, known for producing some of the most powerful AI chips globally, has expressed willingness to collaborate with the US administration on these regulations, while AMD did not provide a statement at the time.

The proposed regulations would exempt a group of 19 allied countries, including the Netherlands and Japan, as well as Taiwan, granting them unlimited access to AI chips. Countries subject to nuclear embargoes, such as China, Russia, Iran, and Venezuela, will continue to be denied access to US-made AI semiconductors.

However, these caps may lead to frustration among certain nations. Geoffrey Gertz, a former White House advisor currently with the Center for a New American Security, noted that global limits on chip access could raise significant concerns among US allies, who may feel uncomfortable with the US acting unilaterally in determining who has access to advanced AI technology.

The US government has indicated it is completing a final review of a regulatory initiative called "Artificial Intelligence Diffusion" drafted by the Commerce Department, suggesting that the announcement may be forthcoming. According to multiple sources, this initiative likely includes the new AI chip caps.

Industry representatives, including the Information Technology Industry Council, express concern that the Biden administration is hastily rolling out a complex rule without sufficient industry input, which could lead to unintended consequences. Naomi Wilson, the council's senior vice president of Asia and global trade policy, has warned that if reports are accurate, the changes will significantly expand the scope of export controls with considerable global ramifications.

The new regulations build on a program announced in September that allows authorized data centers overseas to receive AI chips without requiring a license. To obtain this status, data centers must provide information regarding their customers, operational activities, access controls, and cybersecurity measures.

AI, China, Regulations