Pentagon and Anthropic Leaders To Meet Over Usage Limits on Classified Claude AI Model

Rising friction between artificial intelligence developer Anthropic and the U.S. military is culminating in a critical Tuesday sitdown between company chief Dario Amodei and Defense Secretary Pete Hegseth, according to new reports.The dispute centers on how extensively the Pentagon can deploy Anthropic’s Claude AI systems, with discussions becoming increasingly contentious in recent weeks as both sides maintain firm positions on usage boundaries.A senior Defense Department official talking to Axios said contract discussions have stalled over fundamental disagreements about operational limits. While Anthropic seeks guarantees preventing its technology from powering fully autonomous lethal systems or conducting domestic surveillance, military leaders insist on unrestricted deployment across all legally permissible scenarios.“Our nation requires that our partners be willing to help our warfighters win in any fight,” Pentagon chief spokesman Sean Parnell told NBC News. “Ultimately, this is about our troops and the safety of the American people.”The conflict represents a significant challenge for Anthropic, which currently holds unique status as the sole AI firm with models operating on classified military networks. The company secured a $200 million Pentagon contract last year and has positioned itself as a leader in providing AI solutions to national security agencies.“Anthropic is committed to using frontier AI in support of U.S. national security,” a company representative told NBC on Monday. “We are having productive conversations, in good faith” with military officials about “how to continue that work and get these complex issues right.”The standoff intensified following reports that Anthropic’s systems may have supported the recent operation targeting Venezuelan President Nicolás Maduro. While the exact role Claude played remains unclear, the incident allegedly sparked internal questions at Anthropic about compliance with company policies.Anthropic partnered with defense contractor Palantir in 2024 to deliver AI capabilities on secure government systems. Palantir’s announcement at the time highlighted Claude’s potential for “processing vast amounts of complex data rapidly” and “helping U.S. officials to make more informed decisions in time-sensitive situations.”The underlying tensions reflect broader policy shifts at the Pentagon. Mr. Hegseth’s AI strategy directive in January mandates that all contracts with AI providers eliminate company-imposed restrictions on system usage, allowing “any lawful use” of artificial intelligence technologies. The policy requires implementation within 180 days across all Defense Department AI agreements.Despite the current impasse, Anthropic has consistently supported national security applications of its technology. The company established an advisory council of former defense and intelligence leaders in August 2025 and recently appointed former deputy chief of staff to President Trump, Chris Lidell, to its board.Anthropic, founded in 2021 by former OpenAI researchers, recently completed a $30 billion funding round that valued the startup at $380 billion. Along with OpenAI and Google, Anthropic signed a two-year, $200 million contract with the Defense Department in July 2025 to “prototype frontier AI capabilities that advance U.S. national security.”The Tuesday meeting between Mr. Amodei and Mr. Hegseth represents a critical juncture in resolving the standoff and determining whether Anthropic’s AI systems will continue supporting U.S. military operations under current restrictions or broader terms.

Pentagon and Anthropic Leaders To Meet Over Usage Limits on Classified Claude AI Model

More from the New York Sun

Mission Not Accomplished: Fed Policy Tweaks Under New Chairman Aim To Nip Inflation in the Bud

ICE To Halt Vehicle Stops After Agents Kill Two More Immigrants in Deadly Enforcement Errors: Repor

A Threat Emerges to Wall Street’s Financial Supremacy From Deep in The Heart of Texas

The 21st-Century Twist on Socialism

Steve Marcus Redraws the Visual Narrative of Jewish Pop Art

Can Kevin Warsh Have His Cake and Eat It Too?

Meta Faces Landmark Lawsuit Alleging AI Tools Targeted Disabled Workers for Layoffs

America’s War on Atomic Capabilities and Iran’s Threat To Free Navigation Is About To Escalate