Trump orders all federal agencies to phase out use of Anthropic technology

In a significant move impacting the federal government's adoption of artificial intelligence, the Trump administration has issued an executive order mandating the phase-out of Anthropic's AI technology across all agencies. The order, titled "Securing Federal AI Systems," directs departments to immediately halt new contracts involving Anthropic models and develop transition plans within 90 days to migrate existing systems to alternative AI providers. This decision represents one of the most substantial government interventions in the commercial AI market to date and follows increasing scrutiny of Anthropic's corporate structure and funding sources, which include substantial investment from Amazon and previous backing from entities with ties to effective altruism movements.

The technical implementation requires agencies to identify all systems using Claude models—including the widely deployed Claude 3.5 Sonnet for document processing and Claude Opus for complex reasoning tasks—and replace them with approved alternatives, likely including OpenAI's GPT-4o, Google's Gemini models, or open-source options like Meta's Llama 3. The order cites concerns about "potential foreign influence through algorithmic backdoors" and "undue concentration of critical infrastructure in hands of non-traditional entities." This directive will force hundreds of government AI projects to undergo costly retraining and migration processes, potentially delaying digital transformation initiatives while raising questions about how agencies will maintain functionality during the transition period. The move also sets a precedent for government intervention in AI procurement that could influence private sector adoption patterns.