The Trump administration also directed federal agencies to stop using Anthropic products, while giving the Defense Department a six-month window to phase out the technology, according to Reuters’ account of the order and Pentagon action. In a separate report, The Associated Press reported that Trump posted, “We don’t need it, we don’t want it, and will not do business with them again!”
Anthropic supply chain risk: what the Pentagon designation changes
In procurement terms, a “supply chain risk” label can be used to wall off technology providers that officials believe could compromise sensitive systems, data or operations. The authority the Pentagon cited is tied to federal acquisition rules that allow “covered procurement actions” aimed at reducing supply-chain risk, including limiting disclosure of the underlying rationale in some cases. The legal framework is laid out in 10 U.S.C. § 3252.
What makes this Anthropic supply chain risk designation unusual is that it is typically associated with adversary-linked vendors, not an American AI company that has been selling into government. Reuters reported that the Pentagon has signed $200 million ceiling agreements with multiple major AI labs and warned it could invoke the supply-chain tool if Anthropic did not accept contract language centered on “all-lawful use.”
Practically, the Anthropic supply chain risk label raises immediate questions for defense contractors and their subcontractors about what they can use while performing Pentagon work, and how quickly they would need to replace any embedded systems that rely on Claude or related services.
Why Anthropic says the supply chain risk move is overreach
Anthropic argues the government is stretching a supply-chain mechanism to win a policy dispute over AI safety boundaries. In a company statement, Anthropic said the designation “would both be legally unsound and set a dangerous precedent,” adding: “We will challenge any supply chain risk designation in court.” The statement is posted on Anthropic’s website.
Anthropic also disputes how far the Pentagon can extend the consequences. The company said a supply chain risk determination “can only extend to the use of Claude as part of” Pentagon contracts and “cannot affect how contractors use Claude to serve other customers,” framing the Anthropic supply chain risk fight as a legal question about the scope of federal procurement authority as much as a political one.
At the center of the Anthropic supply chain risk dispute are the “red lines” Anthropic has publicly emphasized: it does not want Claude used for mass domestic surveillance of Americans or to enable fully autonomous weapons without meaningful human responsibility. The Pentagon, for its part, has argued it must retain flexibility to use AI for any lawful mission and has resisted contract language that would lock in private-sector guardrails.
Anthropic supply chain risk fallout for contractors and rivals
The immediate operational impact of the Anthropic supply chain risk designation could land on a sprawling contractor ecosystem. Reuters noted the Pentagon sought assessments of contractor reliance on Anthropic ahead of any designation, underscoring how widely AI tooling can spread through vendors’ internal workflows and products supplied to the government.
The standoff also reshapes competition among AI labs pursuing national security work. The AP report said OpenAI announced a Pentagon deal after Anthropic was punished, and OpenAI CEO Sam Altman wrote that the agreement includes prohibitions on domestic mass surveillance and requirements for human responsibility for use of force. If those terms hold, the Anthropic supply chain risk episode could become a template for how rivals negotiate — or a warning about what happens when talks collapse.
Earlier milestones that set up the Anthropic supply chain risk dispute
While the current showdown is new, the broader relationship between Anthropic and government customers has been building for years — part of why the Anthropic supply chain risk move is landing as a shock across tech and defense circles.
- Amazon deepens its bet (2024): Reuters reported in late 2024 that Amazon invested an additional $4 billion in Anthropic and further positioned AWS as a primary cloud partner — a deal structure that highlighted how intertwined Anthropic had become with major infrastructure players (Reuters report on Amazon’s additional investment).
- A push for federal adoption (2025): In 2025, Reuters reported Anthropic offered Claude to the U.S. government for $1 as AI labs competed for federal approval and contracts — a sign the company was actively courting public-sector work long before the Anthropic supply chain risk fight erupted (Reuters report on the $1 Claude offer).
What happens next in the Anthropic supply chain risk case
Anthropic’s next step is expected to be a court challenge aimed at narrowing or overturning the Anthropic supply chain risk designation, while contractors and agencies track what the Pentagon will require during any phase-out. The company has signaled it will keep pursuing government work within its safety boundaries, but Friday’s actions set up a high-stakes test of whether federal procurement tools can be used to pressure an American AI supplier into changing its policies — or whether courts will curb the government’s reach in the Anthropic supply chain risk showdown.
