Anthropic resumes Pentagon talks after dispute over AI use by the US military: Reports

Artificial intelligence company Anthropic has resumed discussions with the United States Department of Defence over the potential use of its AI technology by the US military.

The talks, which had briefly collapsed last week, could determine whether the Pentagon continues using Anthropic’s AI models or moves to restrict the company’s access to government contracts.

According to people familiar with the discussions, negotiations have restarted in an effort to resolve disagreements around how the technology could be used.

Why talks between Anthropic and the Pentagon broke down

The negotiations were originally led by Anthropic CEO Dario Amodei and Emil Michael, the Pentagon’s under-secretary of defence for research and engineering.

However, discussions reportedly stalled after Anthropic requested guarantees that its AI systems would not be used for two purposes:

Mass surveillance of American citizens

Deployment of fully autonomous weapons

Following the breakdown in talks, US Defence Secretary Pete Hegseth labelled Anthropic a “supply-chain risk”.

This classification is normally used for companies or organisations seen as a potential threat to US national security.

Why the negotiations matter for the AI industry

The dispute has attracted attention across Silicon Valley because it highlights a growing debate over the military use of artificial intelligence.

If both sides reach a new agreement, the Pentagon could continue using Anthropic’s AI tools while reducing concerns about surveillance or weaponisation.

However, failure to reach a deal could lead to Anthropic being restricted from defence-related contracts.

That outcome would also affect competition among major AI companies.

Last week, OpenAI announced that it had reached an agreement allowing the Pentagon to deploy its AI models on classified government networks.

OpenAI CEO Sam Altman said the company was working with defence officials to add safeguards around potential surveillance uses.

A fast-growing AI company under scrutiny

The Pentagon dispute comes at a time when Anthropic is experiencing rapid growth.

According to industry estimates, the company is currently valued at around $380 billion and is projected to generate nearly $20 billion in annual revenue based on current performance.

This would represent more than double the company’s revenue run rate compared with late last year.

Anthropic’s AI assistant app has also recently topped download charts on Apple’s App Store, indicating growing popularity among users.

Silicon Valley rallies behind Anthropic

Several technology groups representing major firms such as Google and Apple have urged the administration of Donald Trump to reconsider designating Anthropic as a national security risk.

Industry groups argue that such a decision could create wider uncertainty for the technology sector, particularly as artificial intelligence companies increasingly work with government agencies.

A final agreement between Anthropic and the Pentagon would help clarify how AI companies collaborate with defence institutions while maintaining ethical boundaries.

For now, the resumed negotiations suggest that both sides are looking for a compromise.