The White House is dictating access to frontier AI models, shifting power from tech giants, sources say

The landscape of artificial intelligence development in the United States is shifting. The Trump administration is now directly dictating access to frontier AI models, a responsibility previously managed exclusively by private technology firms. This change effectively removes the final authority from companies like Anthropic and OpenAI regarding which partners gain access to their most advanced tools.

Historically, labs held the power to vet their own enterprise customers for cybersecurity projects like Project Glasswing and Daybreak. Under new protocols, the White House is launching a clearinghouse program named Gold Eagle. This initiative aims to centralize control over model distribution to address national security concerns. While government officials characterize interactions with labs as voluntary, industry sources indicate that explicit federal approval is now required for model rollouts.

This move comes as the administration attempts to manage the competitive pressure from foreign rivals. Chinese startups are producing powerful models that now rival American technology in performance benchmarks, complicating the government's regulatory strategy. As the administration balances the need for security with the push for innovation, the operational independence of major AI labs faces significant scrutiny and new oversight requirements.