Anthropic’s $200M Pentagon Stand-Off and Safety Policy Shift Heightens Competition for Google AI
Anthropic, awarded up to $200 million in Pentagon AI contracts alongside Google, declined a February 27 deadline to remove restrictions on military use of its Claude system, risking contract cancellation and Defense Production Act invocation. On the same day, it replaced its 2023 binding pause policy with a weaker conditional safety framework, intensifying scrutiny that could affect Google’s defense AI prospects.
1. Pentagon Ultimatum to Anthropic
The Pentagon demanded Anthropic grant unrestricted military access to its Claude AI by 5:01 pm ET on February 27, threatening to cancel its up to $200 million contract, label the company a supply chain risk, and invoke the Defense Production Act if it refused.
2. Anthropic’s Safety Policy U-Turn
Anthropic replaced its 2023 Responsible Scaling Policy’s categorical pause on more powerful AI with a conditional delay framework requiring simultaneous belief of leading the AI race and significant catastrophic risk, a move widely viewed as a rollback of core safety commitments.
3. Context of $200M Pentagon AI Awards
Last July, the Pentagon awarded four companies, including Anthropic and Google, contracts worth up to $200 million each to integrate frontier AI models into classified military networks, marking a key competitive arena for defense AI development.
4. Impact on Google’s Defense AI Prospects
The high-profile stand-off and scrutiny over safety limits on military AI use could shift Pentagon preferences and regulatory attention, potentially affecting Google’s opportunities and valuation in the defense AI market.