The Pentagon has asked Boeing Co (NYSE:BA) and Lockheed Martin Corporation (NYSE:LMT) to produce assessments of their reliance on Anthropic’s Claude artificial intelligence, according to two sources familiar with the exchanges. The outreach represents an initial step toward possibly labeling Anthropic as a "supply chain risk," the sources said.
The supply chain risk label is typically applied to firms tied to adversarial nations, making its consideration for an American AI company notable. The designation process could carry implications for how the Defense Department and its contractors use or limit technologies tied to the flagged firm.
Officials have raised concerns about Anthropic’s resistance to removing or altering certain safeguards designed to prevent Claude from being used in mass surveillance or in the development of autonomous weapons systems. That disagreement has prompted senior defense engagement, including a meeting between Defense Secretary Pete Hegseth and Anthropic CEO Dario Amodei.
Sources said Anthropic was given a deadline on Friday to accept the Pentagon’s conditions or face potential designation as a supply chain risk. Claude is currently the only AI model operating within the military’s classified systems. Earlier in the month the Pentagon was seen engaging other major AI providers to vet alternatives for that role.
Anthropic announced earlier in the week that it was implementing a new safety policy for how it develops its AI models. A separate report indicated that the announced policy change was not connected to the Pentagon’s dispute.
Context and implications
This request to Boeing and Lockheed constitutes a formal inquiry into how defense contractors integrate third-party AI models into classified environments. The outcomes of the assessments could influence contractor procurement decisions and the roster of approved AI technologies used in sensitive defense projects.
What remains limited in public reporting
- The specific contents of the assessments requested from Boeing and Lockheed have not been disclosed.
- There is no public confirmation yet that the Pentagon has issued a supply chain risk designation for Anthropic.
- Details of the Pentagon's alternative engagements with other AI companies have not been specified.