Six weeks ago, the Trump administration slapped Anthropic with a supply-chain risk label and effectively banned its products from the federal government. It was the first time the designation had ever been applied to a domestic company. The message was clear: Anthropic was a national security threat.

Now the same administration wants every major federal agency to use the very model that prompted the ban in the first place.

According to a memo reviewed by Bloomberg, Gregory Barbaccia, the federal chief information officer at the White House Office of Management and Budget, notified officials at the Departments of Defense, Treasury, Commerce, Homeland Security, Justice, and State that OMB is working to establish protections that would allow agencies to deploy Anthropic Mythos in the coming weeks.

From Blacklisted to Essential in 42 Days

The reversal is remarkable even by Washington standards. In early March, Anthropic was deemed too dangerous for government use after a dispute with the Pentagon over how its models could be applied in military operations. Hours before the U.S. began striking Iran, the disagreement escalated into a full-blown ban.

But Mythos changed the calculation. When Anthropic began privately briefing senior government officials on the model’s capabilities before its limited corporate release under Project Glasswing, the response was not caution. It was urgency.

Politico reported earlier this week that the Commerce Department's Center for AI Standards and Innovation had already begun testing Mythos before Anthropic even officially acknowledged the model existed. Staff from at least two other federal agencies had reached out directly to Anthropic requesting access, despite the ban. And at least three congressional committees held classified briefings on the model's implications.

The Model That Scared Central Banks

The government is not alone in scrambling to get access. Over the past two weeks, the Bank of England held urgent discussions with cybersecurity officials after previewing Mythos. The ECB issued warnings to every major European bank. The Bank of Canada called an emergency meeting. Five central banks in two weeks, all reacting to the same model.

The crypto industry has been scrambling separately, with multiple firms lobbying for Mythos access to help counter the threat the model poses to current cryptographic systems. The offensive and defensive applications are so significant that ignoring the model is no longer a viable option for any government serious about cybersecurity.

What Happens Next

The plan is not final. Bloomberg reports that OMB is still establishing the specific protections and guardrails before agencies get access. But the direction is clear: the administration that banned Anthropic is now racing to deploy its most powerful tool.

This tracks with a pattern The AI Post has been documenting for weeks: federal agencies testing Mythos in secret, the government privately begging for access to the company it publicly banned, and two courts reaching opposite conclusions about Anthropic's legal standing. The supply-chain risk label still technically stands. The OMB memo effectively routes around it.

The message from Washington is now impossible to misread: Mythos is too powerful to ban. The question is no longer whether the government will use it, but under what conditions and with what safeguards.

First reported by Bloomberg. Confirmed by Reuters. Additional reporting from Politico and Gizmodo.