The United States government wants American companies to win the global AI race. But a federal judge in Washington asked a different question Tuesday: Does the Pentagon actually want that to happen?
That tension sits at the center of an extraordinary legal confrontation. The Department of Defense designated Anthropic—the company behind the Claude AI model—as a supply-chain risk, a move that would effectively block federal agencies from contracting with one of America's most prominent AI developers. During a district court hearing Tuesday, the presiding judge did not hide his skepticism. He called the DoD's actions potentially "crippling," and questioned whether security concerns were the real motivation.
The DoD's supply-chain designation carries real weight. Federal agencies use it to exclude companies from contracts, grants, and procurement programs. For a company like Anthropic, which has pursued federal partnerships, exclusion means not just lost revenue but diminished relevance in shaping government AI policy. The DoD has not publicly detailed its specific concerns about Anthropic, but the designation alone creates what the judge appeared to recognize as an asymmetric punishment.
Anthropic's lawyers argue the designation lacks substantive justification. The company has contracts with federal customers and has participated in government AI initiatives. Being labeled a supply-chain risk while maintaining those relationships suggests, the company contends, that the real issue is not security but something else entirely.
The DoD has cited national security concerns, echoing language it has used in broader discussions about foreign investment in AI companies. Anthropic has received investment from Chinese-linked entities, which has drawn scrutiny. But critics of the designation—including the judge—suggest that using supply-chain rules to effectively ban a domestic competitor raises questions about whose interests the Pentagon is actually protecting.
The government's position relies on broad discretionary authority. Agencies have long used supply-chain designations without judicial second-guessing. DoD officials have framed AI procurement as a national security imperative, suggesting that limiting exposure to potential foreign influence serves core interests. This argument has bipartisan support in Congress, where legislators have pushed for stricter oversight of AI investments.
The judge pressed harder. If Anthropic poses genuine supply-chain risks, why has the Pentagon continued working with the company on specific projects? The inconsistency suggests the designation may serve a different function—as leverage, as punishment, or as a signal to other AI companies about the costs of certain investment decisions.
What happens next will shape how the government wields procurement power over AI firms. The court will weigh whether the DoD followed its own procedures and whether the designation survives legal challenge. But the deeper question—about whether security policy is being repurposed as industrial policy—extends far beyond this case. If agencies can reshape competitive dynamics in the AI sector through procurement designations, the market for American AI leadership may depend less on who builds the best models and more on who has the right political relationships.