Economy February 27, 2026

White House Orders Immediate Halt to Anthropic AI in Federal Agencies, Six-Month Phase-Out Announced

Administration directs agencies, including the Department of War, to end use of Claude-class tools amid unresolved Pentagon talks over military-use safeguards

By Marcus Reed
White House Orders Immediate Halt to Anthropic AI in Federal Agencies, Six-Month Phase-Out Announced

President Trump directed all federal agencies to stop using technology from Anthropic, setting a six-month window to remove the company’s tools after Anthropic declined to accept limits on military applications. The move follows failed negotiations between the company and Defense Department officials over contractual language and operational constraints.

Key Points

  • President Trump ordered all federal agencies to immediately stop using Anthropic technology and set a six-month phase-out for current users, including the Department of War.
  • Negotiations between Anthropic and the Pentagon remain deadlocked after the company rejected the Pentagon's latest contract language; the company seeks explicit curbs on military use of its AI tools.
  • The dispute centers on two Anthropic demands - no surveillance of U.S. citizens and no autonomous lethal strikes without a human in the loop - and affects defense contracting, classified cloud AI deployments, and government tech procurement.

President Trump on Friday issued a directive requiring all federal agencies to immediately cease using technology from Anthropic, the artificial intelligence company, and ordered a six-month phase-out for agencies that currently rely on the firm's tools. The directive explicitly names the Department of War among those agencies affected and warns Anthropic to cooperate with the transition or face potential civil and criminal penalties.

The directive follows a standoff between Anthropic and Pentagon officials over restrictions governing military use of the company's AI products. Defense and company representatives have been unable to bridge differences on contract language, leaving the parties at an impasse.

Emil Michael, the Under Secretary of Defense for Research and Engineering, described the company's behavior as "unpredictable" in remarks on Bloomberg Television but said defense officials remain open to continued discussions. "So long as they’re in good faith, we’re always open to talks," Michael said. "Up until that deadline, I’m open to more talks and I told them so."

On Thursday, Anthropic rejected the latest offer from the Pentagon, saying the revised contract language did not meet its conditions for limiting military applications of its AI systems. The company has insisted on two specific prohibitions: its tools must not be used for surveillance of U.S. citizens, and they must not be deployed for autonomous lethal strikes without a human decision-maker in the loop.

Those positions have created a direct clash between Anthropic's leadership and senior Defense Department officials. Chief Executive Officer Dario Amodei has held firm on the restrictions, and his stance has produced tensions with Defense Secretary Pete Hegseth, who has pledged to remove what he characterizes as "woke" practices at the agency.

Pentagon officials, for their part, have resisted Anthropic's demand to stipulate the company’s limitations in contract terms. They want to operate Anthropic's Claude software - identified as one of the limited AI tools cleared for classified cloud work - under statutory and regulatory limits but without additional use restrictions imposed by the company.

Responding to pressure from defense officials, Amodei issued a statement on Thursday reiterating the company's refusal to accept the Pentagon's terms. "These threats do not change our position: we cannot in good conscience accede to their request," he said.


The administration's order establishes a six-month timeline for agencies to remove Anthropic systems, signaling a firm federal posture while leaving open the possibility of renewed negotiations before the transition window closes. Until then, Defense Department representatives have indicated they will continue to engage if Anthropic demonstrates good-faith intent to reach an agreement.

Risks

  • Continued impasse could disrupt agencies that currently rely on Anthropic tools, creating operational and procurement challenges in defense and government technology systems.
  • Legal and enforcement risks if Anthropic does not cooperate with the ordered phase-out and the administration follows through on civil and criminal consequences.
  • Uncertainty for classified cloud deployments and defense AI workflows if one of the few tools cleared for such work is removed from federal use, potentially affecting mission support and contractor planning.

More from Economy

Canada posts C$26.14 billion deficit through first nine months of 2025/26 fiscal year Feb 27, 2026 U.S. Signals Possible 'Friendly Takeover' as Cuba Faces Growing Energy Shortage Feb 27, 2026 Nomination Delay Deepens Uncertainty Over Federal Reserve Leadership as Powell’s Term Nears End Feb 27, 2026 Dollar Strengthens After Hotter PPI Print and Rising U.S.-Iran Frictions Feb 27, 2026 NASA Adds In-Orbit Docking Mission to Artemis Roadmap as Lunar Race Intensifies Feb 27, 2026