White House 'Gold Eagle' Program Seizes Control of Frontier AI Access

The Trump administration is moving to centralize control over which companies and entities can access the most powerful AI models from Anthropic and OpenAI, according to CNBC reporting.

The Gold Eagle Program

The newly launched "Gold Eagle" program creates a centralized government body with authority to determine which partners gain access to new AI model rollouts. Under the framework, companies need Washington's sign-off before bringing partners into any model launch.

This directly impacts existing efforts like:

• Anthropic's Project Glasswing — selective rollout of Mythos model to cybersecurity partners
• OpenAI's Daybreak consortium — similar structure for cybersecurity models

From Voluntary to Mandatory

The administration's moves have escalated well beyond the voluntary framework established in a June executive order. The GPT-5.6 release went further — OpenAI agreed to a state-curated access list after federal officials pressed the company to delay.

A White House official maintains that "decisions on timing and scope of releases rest entirely with the companies," but the practical reality is different. CNBC reports the White House is "effectively dictating access to frontier models."

The Anthropic Precedent

Last month, a Commerce Department export control order compelled Anthropic to pull both Fable 5 and Mythos 5 from all users globally over national security concerns. Commerce Secretary Howard Lutnick later restored Mythos 5 access for roughly 100 approved organizations. Fable 5 remains restricted.

Competitive Pressure: Kimi K3

The administration faces pressure from Chinese AI progress. Moonshot AI's Kimi K3 debuted last week, closing in on Fable 5 and GPT-5.6 and surpassing both on at least one independent benchmark. Former AI czar David Sacks called it "concerning," writing: "This is how you lose the AI race."

Key takeaway: The era of companies independently deciding who accesses frontier AI is ending. Government oversight of model deployment is becoming the norm, not the exception.