US Agencies Drop Anthropic, Shift AI Systems to OpenAI

In a sweeping policy shift that could reshape the federal government’s approach to artificial intelligence, several major US agencies are cutting ties with AI firm Anthropic and transitioning to rival company OpenAI. The decision follows a directive from President Donald Trump ordering departments to phase out Anthropic’s technology, including its Claude platform.

According to officials and internal communications reviewed by Reuters, the US Treasury Department, the State Department, and the Federal Housing Finance Agency are terminating their use of Anthropic’s AI products. The move comes after the Pentagon signaled it would classify the company as a supply-chain risk, escalating tensions between the administration and one of the country’s leading AI startups.

Treasury Secretary Scott Bessent confirmed the decision in a post on X, stating that the department is ending all use of Anthropic’s products, including Claude. Similarly, William Pulte, director of the Federal Housing Finance Agency, announced that his agency — along with mortgage giants Fannie Mae and Freddie Mac — will discontinue Anthropic tools.

The State Department is also making a significant change. Its internal chatbot, StateChat, will no longer rely on Anthropic’s models. Instead, it is transitioning to OpenAI’s technology. "For now, StateChat will use GPT4.1 from OpenAI," a department memo said, adding that further information would be available down the line. The State Department has not publicly commented beyond the memo.

President Trump’s directive includes a six-month phase-out period for the Defense Department and other federal agencies currently using Anthropic’s systems. The order represents an unusually forceful federal rebuke of a domestic AI company, particularly one regarded as a key player in advancing US leadership in artificial intelligence technologies critical to national security.

The Pentagon’s reported move to label Anthropic a supply-chain risk marks a significant blow. Such designations are typically reserved for foreign adversaries or vendors seen as security threats. Industry observers suggest the decision could isolate Anthropic within government circles and potentially affect its broader public-sector partnerships.

Meanwhile, OpenAI appears poised to benefit from the transition. Late Friday, the company announced a new agreement to deploy its AI systems within the Defense Department’s classified network. The deal positions OpenAI as a central AI partner for the US government at a time when federal agencies are increasingly integrating advanced AI tools into operations ranging from diplomacy to financial oversight and defense.

The developments highlight intensifying scrutiny of AI providers as Washington weighs issues of security, guardrails, and technological control. As agencies pivot from Anthropic to OpenAI, the federal government’s AI ecosystem is entering a new phase — one defined not only by innovation, but by policy, risk assessment, and political direction.