Judge Blocks Pentagon's 'Illegal' Anthropic Blacklisting

A Legal Win Against the Pentagon

A federal judge on Thursday blocked the Pentagon's blacklisting of Anthropic, delivering a significant victory for the AI company in its escalating dispute with the U.S. military over how artificial intelligence should be governed on the battlefield 12. The ruling emerged from a lawsuit Anthropic filed in California after being designated a national security risk 67.

The root of the conflict traces back to Anthropic's refusal to let its Claude chatbot be used for domestic surveillance or autonomous weapons systems, a stance that led the Pentagon to brand the company a threat 7. The judge's decision described the blacklisting as "illegal and baseless," a sharp rebuke of the government's approach to a company that has positioned itself as a leader in AI safety 7. The case underscores a broader tension between defense agencies eager to harness advanced AI for military applications and AI developers wary of enabling surveillance or weaponization uses that could damage their reputations or violate their own usage policies.

A Company on Multiple Fronts

While the legal battle with the Pentagon draws headlines, Anthropic is simultaneously navigating rapid commercial growth and internal caution over its own technology. The company recently signed a massive $45 billion infrastructure deal with Nscale, a UK-based startup, to secure the computing capacity needed to train and run its Claude models 49. The agreement is the latest sign of Anthropic's aggressive, compute-hungry expansion as it competes for dominance in the generative AI race 4.

That growth has also benefited outside investors. Salesforce reported an 18% jump in its stock price after posting strong AI-driven results, with its stake in Anthropic contributing meaningfully to quarterly profit 8. The investment gain highlights how Anthropic's rising valuation is rippling through corporate balance sheets well beyond its own operations.

At the same time, Anthropic has signaled restraint in its model development. According to its latest risk report, the company does not currently plan to release an internal model referred to as "Model 2," which appears more powerful than its existing top-tier system, even as it acknowledges that AI risks are increasing industry-wide 5. This suggests a deliberate, safety-conscious posture that contrasts with the aggressive infrastructure buildout happening elsewhere in the business.

Competitive and Operational Pressures

Anthropic's rivals are also maneuvering around its practices. OpenAI has reportedly been courting business customers frustrated that Anthropic retains their data, offering a promise not to keep customer data as a competitive differentiator 10. Meanwhile, Anthropic has faced operational hiccups of its own, including a temporary outage affecting multiple Claude models that the company said it has since resolved 3.

Taken together, the coverage portrays a company simultaneously fighting government overreach in court, scaling its infrastructure at enormous cost, drawing investor enthusiasm, and trying to manage the risks of its own increasingly powerful technology.