Pentagon tests AI to replace Anthropic’s Claude for military use
The US Department of Defense (Pentagon) is evaluating alternative AI models to replace Anthropic’s Claude in military workflows. Defense Secretary Pete Hegseth labeled Anthropic a supply-chain risk on Feb. 27, after Anthropic refused to remove safety restrictions tied to mass surveillance and lethal autonomous weapons.
A contract worth up to $200 million (awarded in July 2025) to integrate Claude into classified networks unraveled when Pentagon officials pushed for less-restricted deployment. Public pressure rose in January 2026, with reports that Claude supported intelligence analysis and Iran-related operational planning.
Testing began March 1, 2026, via GenAI.mil, a platform separate from the Pentagon’s Maven Smart System. Twenty-five designated military personnel are running evaluations across workflows previously relying on Anthropic’s Claude.
Competing models under review include OpenAI, Google, and Grok from Elon Musk’s xAI. Contractors using Claude were given six months to find alternatives. Anthropic has also filed lawsuits over the decision.
For traders, the key takeaway is not direct market linkage, but the defense AI procurement signal: US government demand may shift away from tightly safety-gated deployments like Anthropic’s Claude, while favoring broader “deployability” models for classified tasks. This could influence broader AI-sector sentiment and risk appetite around AI-related tech narratives, even if crypto flows are not directly implicated.
Neutral
This is a defense AI procurement and policy decision, not a crypto-native catalyst. It may affect broader AI-sector sentiment, but there’s no direct linkage to major tokens’ fundamentals.
Historically, government or defense procurement shifts (especially around “deployability” vs “safety constraints”) tend to create short-term narrative swings in AI/tech equities and related sentiment, without necessarily translating into immediate, measurable on-chain or token demand. Here, the key timeline (Feb. 27 supply-chain risk label; March 1 start of GenAI.mil testing; up to $200M Claude contract unwind) is clear, but it doesn’t introduce new crypto regulation, major protocol changes, or token-specific partnerships.
Short-term: likely neutral for crypto prices; traders may watch for secondary effects on AI-tech risk appetite rather than expect direct BTC/ETH movement.
Long-term: continued reshaping of defense AI vendors could shift which AI products are considered “enterprise/classified ready,” impacting tech-sector expectations. That can indirectly influence liquidity conditions (risk-on/risk-off) for the whole market, but the effect on crypto is expected to be indirect and gradual.