U.S. Agencies Ditch Anthropic, Tap OpenAI After Trump Order

Washington is seeing a coordinated shift in federal AI partnerships after a White House directive ordered the phase-out of Anthropic’s tools. The Treasury Department, the State Department and the Federal Housing Finance Agency have terminated all use of Anthropic products, including the Claude model, as Washington seeks to recalibrate its AI infrastructure in line with national security and governance priorities. The move follows a public order from the president directing government agencies to end work with Anthropic, with the Pentagon signaling it will label the company a supply-chain risk in the broader effort to control critical technology.

In a related development, the State Department announced it will switch its in-house chatbot, StateChat, from Anthropic to OpenAI. The current plan uses GPT-4.1 from OpenAI, with officials promising further information as the transition proceeds. The State Department, along with other agencies, is planning a six-month phase-out window for Anthropic’s products, part of a wider realignment that touches multiple parts of the government.

The shift extends beyond these immediate changes, as rivals to Anthropic, notably OpenAI, have been expanding deployments in government networks. OpenAI recently disclosed a deal to deploy its technology within the Defense Department’s classified network, a move that underscores a broader pivot toward AI partners that can meet security and oversight requirements. The transition signals a significant reorientation in how the federal government weighs AI vendors, security guardrails, and the balance between innovation and risk.

Industry observers say the move could reshape federal AI partnerships for years to come, potentially influencing how other agencies evaluate vendors and implement guardrails. Officials did not immediately provide remarks, and details on timelines and oversight mechanisms are expected to follow as the phase-out progresses.

Overall, the changes mark a definitive stance in the government’s approach to AI tools, prioritizing certain vendors perceived to align with current security, governance, and procurement standards.

#AnthropicExit #OpenAI #AIPolicy