On January 13, the federal government proposed a global policy to limit the distribution of powerful American-made GPUs, in an effort to promote American dominance over China in the AI chip industry.
“To enhance America's national security and economic strength, it is essential that we do not offshore this critical technology and that the world's AI runs on American rails,” the White House announcement reads.
The White House moved forward with the proposal despite pushback from several tech giants, including NVIDIA and Oracle. If the rule is implemented, it will be enforceable after 120 days, following a comment period. The incoming administration will determine whether the rule will go into effect.
Proposed rule classifies countries into allied, restricted, or limited categories
The Interim Final Rule on Diffusion of Artificial Intelligence proposes restrictions on the distribution of US chips to ensure US market share against competing technologies in China. Issued under the Department of Commerce's Bureau of Industry and Security, the rule mitigates national security risks exacerbated by generative AI, including cyberattacks.
The rule has six main parts:
- 18 US allies will have unlimited access to chip sales.
- “Orders for chips with collective computing power of up to approximately 1,700 advanced GPUs” are exempt, the White House said. Therefore, most universities and medical and research institutions can order chips without obstacles.
- Trusted partners known as “Universal Verified End Users” are permitted by the US to place a small percentage of their global AI computing capacity globally.
- Countries that are not on the US “Countries of Concern” list or the list of close allies that meet certain security requirements may be listed as “Verified Domestic End Users.” This designation allows them to purchase computing power equivalent to up to 320,000 advanced GPUs from the US for use in neutral countries over the next two years. (The “countries of concern” are China and its Special Administrative Regions, Hong Kong and Macau.)
- Foreign governments, healthcare providers, and companies outside the US or their close allies that are not designated verified domestic end users can purchase up to the equivalent of 50,000 US Advanced GPUs per country.
- Governments that sign agreements with the United States to specify export control, clean energy and technological security measures can double the number of chips they have been allocated.
China will have restricted access to advanced US technology in general and to the weights of basic AI models. Russia also remains blocked from purchasing advanced chips from the US.
Which countries have unlimited access to US AI chips?
The 18 allied countries with unlimited access to US AI chips under this policy are:
- Australia.
- Belgium.
- Canada.
- Denmark.
- Finland.
- France.
- Germany.
- Ireland.
- Italy.
- Japan.
- Netherlands.
- New Zealand.
- Norway.
- Republic of Korea (South Korea).
- Spain.
- Sweden.
- Taiwan.
- United Kingdom.
As CNN notes, categories of countries with limits on the amount of American chip technology they can access could help prevent China from acquiring American-made chips or gaining a foothold in the high-tech industry in those countries. Israel and Mexico are among the countries with limited access.
Depending on the size of their operations, companies outside the US could experience delays in their supply chains or limited opportunities to add AI features to their products due to this policy.
“While the administration's decisions and export controls produce limited short-term advantage, they may produce a broader, long-term loss for American technology leadership,” Benjamin Lee, professor of engineering and technology, told TechRepublic in an email. computer science from the University of Pennsylvania. .
“In the short term, export controls will slow the deployment in some countries of the most advanced GPUs and largest AI data centers. But in the long term, export controls will cause other countries to develop their own hardware architectures or software models.”
NVIDIA and Oracle strongly oppose the measure
NVIDIA Vice President of Government Affairs Ned Finkle disagrees with the new rules on two main points. He says distributing powerful AI chips is a matter of national security and that restricting the purchase of AI in some countries will benefit the US.
“Although disguised as an 'anti-China' measure, these rules would do nothing to improve US security,” Finkle wrote. “The new rules would control technology around the world, including technology that is already widely available in major gaming PCs and consumer hardware.”
SEE: The EU approved a major merger between simulation software company Ansys and chip design software provider Synopsys.
Ken Glueck, executive vice president at Oracle, acknowledged the importance of some restrictions around the use of AI in cases such as weapons of mass destruction and superintelligent AI. However, he also opposes the new White House policy, calling the rule “very complex and tremendously broad” in a Jan. 5 blog post.
“In a single confusing move, BIS retroactively regulates global cloud GPU deployments; reduces the global market for US chip and cloud suppliers; sets volume restrictions; “It tells 20 countries that they can only be trusted if they accept new unilaterally imposed terms, including semi-annual reporting and certification requirements, and likely pushes the rest of the world to adopt Chinese technology,” Glueck wrote.
However, in a statement to the Associated Press, Microsoft President Brad Smith was less critical: “We are confident that we can fully meet the high security standards of this rule and meet the technological needs of countries and customers around the world who trust us. .”
