In one of the Biden-Harris Administration’s last moves during its final weeks in office, it issued the “Export Control Framework for Artificial Intelligence (AI) Diffusion” as an Interim Final Rule (IFR). The nearly 200-page document aims to protect the US from AI-related national security risks stemming from its dual-use nature and rapid advancements. Nvidia described the export controls as “misguided.”
The key element of the rule is implementing country-specific caps on the computation power of GPU exports, with 18 trusted countries exempt from the law and 140 not exempt, to control its global reach. The Information Technology & Innovation Foundation (ITIF) has voiced concerns that AI diffusion rule poses significant compliance challenges, as the stringent caps would be nearly impossible for the US government to track and oversee.
Meanwhile, the US Secretary of Commerce, Gina Raimondo, has stated in the administration’s press release that “this policy will help build a trusted technology ecosystem around the world and allow us to protect against the national security risks associated with AI while ensuring controls do not stifle innovation or US technological leadership.” Despite a 120-day comment period, the rule may take effect anytime between Friday, January 17, and the end of this comment period.
The deep complexity of the AI Diffusion framework makes global trade compliance more critical than ever for US semiconductor exporters and buyers. Correct Rules of Origin will likely be vital in determining whether GPU chips consist of US-origin components, and exporters need efficient measures to ensure compliant shipments and avoid penalties or delays.
The White House explains in its factsheet that the AI Diffusion rules strive to improve previous chip controls, like the CHIPS and Science Act, by “thwarting smuggling, closing other loopholes, and raising AI security standards.”
The key goals of AI diffusion include:
Preventing countries of concern from training AI models,
Controlling the transfer of advanced models to non-trusted entities, and
Protecting the data of advanced models by storing them securely while preventing illicit applications.
The Biden-Harris administration aims to prevent AI-related national security risks like building weapons of mass destruction, cybercrime, and mass surveillance.
The White House’s Mechanisms for AI Diffusion:
- Eighteen allies and partners with US-aligned technology protection policies will not be subject to export controls.
- GPU shipments will not require a license or count towards the caps if their computation power equals 1,700 of the most advanced GPUs, currently Nvidia’s H100 Tensor Core GPU.
- Trusted, high-security entities can obtain “Universal Verified End User” (UVEU) status. With this status, entities can host up to “7% of their global AI computational capacity” in other countries.
- Entities that meet the US’s security requirements with headquarters in any location that is not of concern to the US can obtain a “National Verified End User” (NVEU) status. This would enable them to “purchase computational power equivalent to up to 320,000 advanced GPUs over the next two years.”
- Entities without VEU status outside of closely allied countries can purchase “up to the equivalent of 50,000 advanced GPUs per country.”
- Governments can double their GPU caps up to the equivalent of 100,000 advanced chips if they sign arrangements to ensure shared values surrounding AI and align their efforts for export controls, clean energy, and technology security with those of the US.
Nvidia’s Vice President of Government Affairs, Ned Finkle, believes that “America wins through innovation, competition, and by sharing our technologies with the world — not by retreating behind a wall of government overreach.”
“The new rule risks causing unintended and lasting damage to America’s economy and global competitiveness in semiconductors and AI by ceding strategic markets to our competitors,” shared The Semiconductor Industry Association. “We stand ready to work with leaders in Washington to chart a path forward that protects national security while allowing us to do what America does best – compete and win globally.”
Ken Glueck, Executive Vice President of Oracle, stated on January 5 that “the Biden Administration’s “Export Control Framework for Artificial Intelligence Diffusion” will go down as one of the most destructive to ever hit the US technology industry.”
He continues, “We all agree on the need to protect national security from the very real threats from certain AI uses; however, this rule does more to achieve extreme regulatory overreach than protect US interests and those of our partners and allies. It practically enshrines the law of intended consequences and will cost the US critical technology leadership”.
While Microsoft has not commented directly on the framework, the Vice Chair and President of Microsoft, Brad Smith, stated on January 3, “The best response for the United States is not to complain about the competition but to ensure we win the race ahead. This will require that we move quickly and effectively to promote American AI as a superior alternative. And it will need the involvement and support of American allies and friends.”
Smith continues, “We are designing AI technology that protects cybersecurity, privacy, digital safety, and other responsible uses of AI. And we are making this technology available around the world through datacenters that meet the U.S. Government’s highest cyber and physical security standards.” These statements suggest that Microsoft may not adopt a hostile opinion on the framework.
Trade Compliance For Seamless AI Diffusion
Understanding US allies, embargoes, and trade relations will be necessary to assess your company’s position as an exporter or purchaser of US AI chips. Knowing whether your company or supply chain will be affected by the AI Diffusion export controls will help you establish what new licenses you can benefit from.
With Trump’s administration nearing inauguration day, it will inevitably have the final say over the longevity of this new rule. Whether Trump and his administration see the regulations as supporting their protectionist policies or burdening America’s economy while advancing China as an AI leader is yet to be confirmed. However, the new administration will strongly influence the future of semiconductor trade with the US by enforcing or vetoing the framework alongside implementing new tariffs.
The New York Times reports that the framework could encourage companies to build data centers in the US over regions like the Middle East, which could attract tech firms and create work for construction companies. Trump has reportedly “expressed support” for building data center infrastructure in America, which could influence his final decision on the framework.
With uncertain regulations and the potential for complicated country-specific chip caps that are challenging to track, partnering with a trade compliance expert like TecEx is essential. Not only can we help you navigate intricate compliance requirements, obtain necessary licenses, and ensure an airtight Certificate of Origin, but we can also take on your end-to-end shipping risk.
TecEx is your trusted partner in successful global GPU shipments with our existing understanding and vast experience in safely shipping AI tech.