Policy developments around data center regulations, AI governance, and chip export controls alongside OpenAI discussions.

With one week left in office, President Biden signed an executive order setting aside federal lands for artificial intelligence data centers, with the full cost borne by developers of AI foundation models like OpenAI's GPT-4o. The order requires AI model developers to ensure clean energy sources power these data centers, addressing the notorious energy consumption of intense AI workloads. This executive order follows Biden's October 2023 order establishing guardrails for frontier or foundation AI models, which included requirements for government evaluation of AI systems before deployment in areas such as cybersecurity and national security. Biden also pledged to develop labeling and content provenance mechanisms so consumers can identify AI-generated content.

The regulatory approach reflects divergent international strategies. The EU AI Act employs a three-tier risk framework assessing applications as unacceptable risk (such as government-run scoring of individuals based on social standing), high-risk (like resume-scanning AI tools that rank job applicants), or neither banned nor high-risk. Meanwhile, U.K. Prime Minister Keir Starmer announced on Monday, January 13, an action plan to make Britain a leader in AI, including expanding data center capacity for AI workloads, with formal regulations coming. His predecessor Rishi Sunak had unveiled an AI regulatory framework for existing regulators to follow.

Biden also expanded chip export controls originally implemented in 2022 and 2023 to prevent China and other adversary nations from accessing AI hardware. The new regulations divide the world into groups: 18 allies and partners receive no restrictions at all, while universities and research organizations can purchase up to 1,700 advanced GPUs. However, more than 120 other countries face new restrictions on establishing AI computing facilities. Trusted entities based in allied countries have more flexibility, while non-allied countries can acquire up to 50,000 advanced GPUs per country. Additional rules keep AI model weights secret from untrusted entities and impose other security controls.

The restrictions carry significant implications for Nvidia, which commands more than 80% market share in GPU chips used for training and inference of AI models. Ned Finkle, Nvidia's vice president of government affairs, criticized the regulations as "misguided" policy, characterizing Biden's expanded export controls as a "200+ page regulatory morass, drafted in secret and without proper legislative review." Finkle warned that such regulatory action "threatens to squander America's hard-won technological advantage" and expressed hope for a return to the first Trump administration's policies fostering AI innovation. The Semiconductor Industry Association echoed concern, stating: "We're deeply disappointed that a policy shift of this magnitude and impact is being rushed out the door days before a presidential transition and without any meaningful input from industry."

Meanwhile, as OpenAI CEO Sam Altman signaled plans to attend President Trump's inauguration, OpenAI released its "AI in America" economic plan. The blueprint asserts: "We believe America needs to act now to maximize AI's possibilities while minimizing its harms. AI is too powerful a technology to be led and shaped by autocrats, but that is the growing risk we face, while the economic opportunity AI presents is too compelling to forfeit."