A federal judge in California issued an indefinite block on March 26 to stop the Pentagon from enforcing its supply chain risk designation against Anthropic, the company behind the Claude AI models. Judge Lin ruled the designation was “classic illegal First Amendment retaliation” — a finding that carries significant implications for every business building on commercial AI platforms.

This is one of the most consequential AI governance rulings to date, and business leaders evaluating or deploying AI need to understand what happened, why it matters, and what it does not mean.

What Happened

In February, Defense Secretary Pete Hegseth designated Anthropic a supply chain risk — a label typically reserved for companies with ties to foreign adversaries like China or Russia. The designation was unprecedented in its application to a domestic AI company. It effectively barred federal agencies from using Anthropic products and required any company doing defense work to prove it did not use Anthropic’s technology.

The origin of the dispute was a contract negotiation. Anthropic had signed a $200 million contract with the Pentagon in July 2025. When the parties tried to expand Claude’s deployment on the DOD’s GenAI.mil platform, talks broke down over two conditions Anthropic wanted included: that its AI would not be used for mass surveillance of U.S. citizens, and that it would not power fully autonomous weapons. The Pentagon wanted unrestricted access for “all lawful purposes.”

Negotiations collapsed. Hegseth then issued the supply chain risk designation. The General Services Administration terminated Anthropic’s OneGov contract, cutting off Claude access to all three branches of the federal government.

Anthropic filed two separate lawsuits. One in California federal district court, one in the U.S. Court of Appeals in Washington. On March 26, Judge Lin — a Biden appointee — sided with Anthropic.

Her written ruling was direct: “The Department of War’s records show that it designated Anthropic as a supply chain risk because of its ‘hostile manner through the press.’ Punishing Anthropic for bringing public scrutiny to the government’s contracting position is classic illegal First Amendment retaliation.”

The judge found Anthropic likely to succeed on its First Amendment and due process claims. The block takes effect in one week to allow the government to appeal.

The Industry Reaction

The support for Anthropic across the industry was notable. Dozens of scientists and researchers from OpenAI and Google DeepMind — Anthropic’s direct competitors — filed an amicus brief in their personal capacities, arguing the supply chain designation could damage U.S. AI competitiveness and chill public discussion about AI risks. Microsoft filed a supporting brief as well. So did retired U.S. military leaders and various industry trade groups.

The breadth of support signals that the AI industry views government retaliation against vendors over policy disagreements as an existential concern — not just an Anthropic problem.

Anthropic’s response to the ruling: “We’re grateful to the court for moving swiftly, and pleased they agree Anthropic is likely to succeed on the merits.”

The administration’s response, from Under Secretary of Defense Emil Michael: “There are dozens of factual errors in the 42 page judgment rushed out in 48 hours during a time of conflict.”

The legal fight is not over. The Justice Department can seek an emergency stay from the Ninth Circuit.

What This Means for Business

For commercial Anthropic customers, this changes nothing operationally. Claude has remained fully available to non-government customers throughout this dispute. Google and Microsoft have confirmed Anthropic’s tools continue to operate normally outside defense contexts. If you are using Claude-powered tools today, you will continue using them.

For enterprise AI buyers, the story is more interesting. This dispute has surfaced a set of questions that most vendor evaluation processes do not ask: What are this vendor’s ethical limits? What will they refuse to build? What happens if a government tries to cut them off? What is our continuity plan if a vendor becomes unavailable?

These are not hypothetical concerns. They happened. For a period, federal agencies lost access to an AI platform they had contracted for. The situation was resolved, but the vulnerability was real.

For businesses evaluating which AI vendors to build on: Vendor concentration risk has always been a factor in technology strategy. AI platforms add a new dimension — political and regulatory risk. A vendor’s stance on government use cases, weapons development, or surveillance can now create ripple effects that reach commercial customers. That does not mean avoid Anthropic. It means understand the full risk profile of your AI stack.

For anyone building in regulated industries: The supply chain risk designation mechanism, even when blocked by courts, signals that government agencies are looking at AI vendor relationships differently than they look at traditional software vendors. Healthcare, finance, defense contractors, and any organization with government contracts should have documented visibility into which AI models their systems use and under what terms.

The Bigger Principle at Stake

Beyond the specific facts of this case, what is at stake is a question about the relationship between government and the companies building AI infrastructure: can the government penalize a vendor for refusing to build weapons or surveillance tools?

Judge Lin’s ruling says no, at least not through the supply chain risk mechanism without proper process. The ruling is preliminary, not final. But the framing matters.

If Anthropic had complied with Pentagon demands and agreed to build autonomous weapons capability into Claude, would enterprises in healthcare, education, and financial services feel comfortable using those same models? Almost certainly not. The ethical limits Anthropic drew are, paradoxically, part of what makes their technology trustworthy for commercial use.

That tension between government capability demands and commercial trustworthiness is not going away. It will play out again with other vendors, other countries, and different compliance demands.

---

If you’re deciding where to start with agents, start here. The free Working With Claude field guide walks through the ecosystem, Claude Code, and a real rollout plan. Get your copy.