Judge blocks Pentagon move against Anthropic

A federal judge temporarily blocked the Pentagon from labeling Anthropic a supply chain risk after finding major gaps between public threats, legal authority, and the government’s courtroom arguments. The dispute has become a test of how far the government can go in punishing an Artificial Intelligence company over political and contractual conflict.

A California judge temporarily blocked the Pentagon from labeling Anthropic a supply chain risk and ordering government agencies to stop using its Artificial Intelligence. The order marked the latest turn in a month-long conflict that is still unresolved, with the government given seven days to appeal and a separate Anthropic case in Washington, DC, still pending. For now, Anthropic remains shut out of government business even as the most aggressive enforcement steps have been paused.

Judge Rita Lin’s 43-page opinion framed the fight less as a national security emergency than as a contract dispute that escalated because the government bypassed ordinary process. Court documents said the government used Anthropic’s Claude for much of 2025 without complaint, including through Palantir, where defense employees accepted a government-specific usage policy. Jared Kaplan said that policy “prohibited mass surveillance of Americans and lethal autonomous warfare.” The conflict intensified only when the government sought to contract with Anthropic directly.

The ruling focused heavily on the mismatch between public statements from senior officials and the legal steps required to impose a supply chain risk designation. President Trump’s Truth Social post on February 27 attacked Anthropic as “Leftwing nutjobs” and directed federal agencies to stop using the company’s Artificial Intelligence. Pete Hegseth then said he would direct the Pentagon to label Anthropic a supply chain risk. The judge found that the required actions for that designation were not completed. Letters to congressional committees said less drastic alternatives had been considered and rejected, but offered no detail. The government also claimed Anthropic could implement a “kill switch,” but its lawyers later admitted they had no evidence of that.

The court also zeroed in on Hegseth’s statement that “No contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.” Government lawyers later conceded that the Secretary lacked that power and agreed the statement had “absolutely no legal effect at all.” Those contradictions helped support Anthropic’s claim that the government was retaliating against the company for its views and conduct rather than following lawful procurement procedures. The judge wrote that officials had sought to publicly punish Anthropic for its “ideology,” “rhetoric,” and “arrogance.”

Even with the temporary injunction, the pressure on Anthropic is unlikely to end. The government is expected to appeal, and officials may still have other ways to discourage defense contractors from working with the company. Charlie Bullock said, “I think it’s safe to say that there are mechanisms the government can use to apply some degree of pressure without breaking the law,” while noting that the outcome depends on how committed the government is to punishing Anthropic. The dispute also exposed a contradiction inside the administration’s approach: Claude was important enough to government operations that Trump said the Pentagon needed six months to stop using it, even as senior officials pursued an aggressive campaign against the company.

72

Impact Score

Anumana wins FDA clearance for pulmonary hypertension ECG Artificial Intelligence tool

Anumana has received FDA 510(k) clearance for an Artificial Intelligence-enabled pulmonary hypertension algorithm designed for use with standard 12-lead electrocardiograms. The company says the software can help clinicians spot early signs of disease within existing workflows and without moving patient data outside the health system environment.

Anu Bradford on tech sovereignty and regulatory fragmentation

Anu Bradford argues that Europe is wavering in its role as the world’s digital rule-setter just as governments everywhere move toward more state control over technology. Global companies are being pushed to treat geopolitical risk, data sovereignty, and Artificial Intelligence governance as core strategic issues.

Mistral launches text-to-speech model

Mistral has expanded its Voxtral family with a text-to-speech system aimed at enterprise voice applications. The company is positioning the open-weights model as a flexible alternative for organizations that want more control over deployment, cost and customization.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.