Pentagon summons Anthropic CEO over military use of Claude

US Defense Secretary Pete Hegseth will meet with Anthropic CEO Dario Amodei at the Pentagon on February 24 amid ongoing disputes over military access to Anthropic’s AI model, Claude.

Sources indicate the meeting will be tense, with the Pentagon seeking broader use of Claude for classified defense and intelligence purposes.

Anthropic has expressed willingness to relax some restrictions but opposes using its AI for mass surveillance of Americans and autonomous weapons.

The Pentagon has threatened to label Anthropic a “supply chain risk,” which could void existing contracts and impact other contractors.

Negotiations have reportedly shown little progress, and officials suggest the Pentagon may present an ultimatum.

🔗 Source: Axios

🧠 Food for thought

Implications, context, and why it matters.

Anthropic’s growth and the Pentagon’s other options shape the stakes

The dispute involves a two-year, $200 million agreement awarded to Anthropic in July 2025 to prototype AI for national security 1.
Anthropic has room to push back. Its revenue run-rate rose from about $1 billion at the start of 2025 to more than $5 billion by August 2025, backing claims of greater financial independence 2.
The Department of Defense (DoD) has other vendors. That same day, it also granted contracts of up to $200 million to Google, OpenAI, and xAI, widening the pool of suppliers 3.
The disagreement hits operations because Anthropic says Claude is built into mission workflows on classified networks through partners like Palantir (a data analytics and defense software company). Reports also say Claude is the only AI system approved on the Pentagon’s classified networks, so limits on use can affect day-to-day work inside DoD systems 1.

A red line dispute is emerging over AI use in warfare and surveillance

The clash pushes the tech industry past broad “responsible AI” statements and toward clear lines for military and intelligence use, including surveillance and weapons-related work 4.
A “supply chain risk” label from the Pentagon could force U.S. defense contractors to stop working with Anthropic, which would also squeeze firms that sell into the defense market 5.
The result could split the AI market. Elon Musk’s xAI is building “Grok for Government” for public-sector customers, while other providers keep tighter limits on certain uses 3.
This marks a shift as the Pentagon seeks AI tools that work for all “lawful purposes,” including weapons development and intelligence operations, raising pressure on companies’ self-set guardrails 4.