Pentagon clashes with Anthropic over AI safeguards as $200 million contract stalls

Pentagon and Anthropic Reach Standstill Over Military AI Use

The Pentagon and Anthropic are locked in a dispute over AI safeguards that could reshape how Silicon Valley companies influence military technology deployment. After weeks of contract talks under a deal worth up to $200 million, the U.S. Department of Defense and the San Francisco-based AI developer have reached an impasse, according to Reuters 2

2

. Six people familiar with the matter confirmed the standstill, revealing that the company's position on usage policies has intensified disagreements with the Trump Administration 3

3

.

The conflict centers on whether Anthropic must eliminate safeguards built into its Claude LLM models that specifically prevent deployment in scenarios involving autonomous weapons targeting and domestic surveillance. In discussions with government officials, Anthropic representatives raised concerns that its tools could be used to spy on Americans or assist weapons targeting without sufficient human oversight 3

3

. The company maintains these safeguards are a key part of its ethical AI standards and usage policies.

Pentagon Officials Push to Deploy Commercial AI Without Restrictions

Pentagon officials, now operating under the Trump administration's renamed Department of War, have bristled at Anthropic's guidelines. Citing a January 9 department memo on AI strategy, Pentagon officials argue they should be able to deploy commercial AI technology regardless of companies' usage policies, as long as they comply with U.S. law 3

3

. This position directly challenges Silicon Valley's influence over how intelligence agencies and military personnel use increasingly powerful AI on the battlefield.

Yet Pentagon officials would likely need Anthropic's cooperation moving forward. The company's models are trained to avoid taking steps that might lead to harm, and Anthropic staffers would be the ones to retool its AI for military applications 3

3

. An Anthropic spokesperson stated that its AI is "extensively used for national security missions by the U.S. government" and that the company is "in productive discussions with the Department of War about ways to continue that work" 3

3

.

CEO Dario Amodei Warns Against Becoming Like Autocratic Adversaries

AnthropicCEO Dario Amodei addressed the ethical concerns in a blog post this week, warning that AI should support national defense "in all ways except those which would make us more like our autocratic adversaries" 3

3

. Amodei, who was among Anthropic's co-founders critical of recent fatal shootings of U.S. citizens protesting immigration enforcement actions in Minneapolis, described the incidents as a "horror" in a post on X 3

3

. These deaths have compounded concern among some in Silicon Valley about government use of their tools for potential violence.

High Stakes for Anthropic's National Security Business

The spat threatens Anthropic's Pentagon business at a delicate time for the Amazon-backed company 5

5

. The startup is preparing for an eventual public offering while having spent significant resources courting U.S. national security business and seeking an active role in shaping government AI policy 3

3

. Anthropic is one of a few major AI developers awarded contracts by the Pentagon last year, alongside Google, OpenAI, and Elon Musk's xAI 3

3

.

The company has boosted its 2026 revenue projection by 20 percent, now expecting sales to reach $18 billion this year and $55 billion in 2027 5

5

. Claude achieved a $1 billion run rate revenue just six months after its public launch, and Anthropic recently completed a funding round valuing the company at $350 billion 5

5

.

Given Anthropic's history of partnering with Lawrence Livermore National Laboratory and Palantir for classified security and intelligence work, observers question whether the company anticipated these demands from the Department of War 1

1

. Anthropic already has a special version of Claude for "US national security customers" 1

1

. The discussions represent an early test case for whether Silicon Valley can maintain ethical concerns while pursuing lucrative government contracts, and if the company doesn't accept the terms, other AI companies may rush in to capture the Pentagon's substantial budget 1

1

.