The White House is drafting a directive that would allow agencies to bypass a previous supply chain risk designation against Anthropic, paving the way for the company’s most powerful AI model, Mythos, to be used by the federal government, according to sources familiar with the matter.
The move marks a dramatic reversal for an administration that earlier this year claimed the firm was such a severe security threat that it had to be removed from government systems entirely. The potential executive action, still in draft form, could also include broader steps related to the government’s use of AI.
A source described White House efforts as a way to “save face and bring em back in.” Earlier this month, Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent met with Anthropic CEO Dario Amodei in what both sides called a productive introductory discussion on collaboration.
The administration is convening companies across several sectors this week to shape the potential action and best practices for deploying Mythos. Those gatherings include “table reads” of possible guidance, according to insiders.
The sudden rapprochement highlights the tension between security concerns and the desire to leverage cutting-edge AI for government operations. Critics may question whether the reversal is driven by political expediency rather than a genuine reassessment of risk.