In March 2026, the Pentagon blocked Anthropic — a company that makes AI software called Claude — from selling to federal agencies. On June 9, the Trump administration said this was not punishment; it was a security decision. Anthropic disagrees and is taking the matter to court.

The designation at stake is called a "supply-chain risk." In practical terms, it means that government contractors cannot use Anthropic's AI models. Since major federal projects often build on vendor tools, anyone with a government contract that wanted to use Claude would face a choice: switch to a different AI system or request special permission. Making that switch takes months and costs money.

For Anthropic, which counts government agencies and contractors as paying customers, the blacklisting cuts off a significant revenue stream.

The Rules Are Getting Looser, But It is Not Over

By early June, news reports suggested tensions were easing. Anthropic was preparing to go public — to sell stock to the general public for the first time — and operating under a government blacklisting creates problems. Investors worry about whether the company will keep earning revenue if the government keeps blocking it. Those worries lower the price investors will pay for the stock.

Even so, easing tensions do not settle the legal fight. The administration did not cancel the blacklisting; it simply argued that Anthropic's lawsuit has no legal grounds. This leaves Anthropic in a strange position: the government might quietly let contracts go through, but the official restriction stays on the books.

Fighting on a Second Front

Anthropic is not just suing. On June 10, the company urged Congress to pass federal AI safety rules — rules that states cannot override. Several states have started writing their own AI rules, and Anthropic argues that a federal standard is better than letting each state do its own thing.

This position serves Anthropic's interests. The company has spent years building a reputation for safe, well-tested AI, and federal safety rules would reward that track record. Smaller competitors without as much safety infrastructure would struggle more to comply. But Anthropic's argument also makes a practical point: having no national standard at all leaves companies and consumers more confused, not less.

By fighting the blacklisting in court and pushing for federal rules in Congress, Anthropic is betting that both moves are necessary for its future. Winning in court restores government business now. Federal safety rules, if they match what Anthropic already does, would give the company a lasting competitive edge.

What Matters for Business Right Now

If you work at a company with government contracts and are thinking about using Anthropic's AI, the safe move is to wait. The blacklisting is still official until a judge or the Pentagon formally cancels it. Using Anthropic's tools while the blacklisting is active could create legal problems.

The Congressional push for federal AI rules moves more slowly. But if Congress passes a law setting safety standards for the most powerful AI models, every company that uses cutting-edge AI in regulated industries will have to adapt. What happens in the next few months in Congress will shape that landscape.

This situation — a company fighting the government on multiple fronts at once while trying to grow its business — is not new in the tech industry. The specific AI technology is new, but the structural pressure is familiar.