The Department of Defense announced May 1 that it had finalized agreements with eight commercial AI companies to deploy their technologies on the department's most sensitive classified networks. The eight firms are Amazon Web Services, Google, Microsoft, NVIDIA, OpenAI, SpaceX, and Reflection, with Oracle added hours after the initial announcement. The agreements authorize those systems to operate on networks classified at Impact Level 6, which covers secret-level data, and Impact Level 7, the department's designation for its most highly classified systems. The department described the purpose as deploying advanced AI capabilities "for lawful operational use." Anthropic, whose Claude model was previously the only AI integrated into the department's classified infrastructure, was not included. [POLITICO]
The legal and statutory basis for the supply chain exclusion underpinning Anthropic's absence is specific. The Department of War has invoked two authorities to support its supply chain risk designation against Anthropic: 10 U.S.C. § 3252 and the Federal Acquisition Supply Chain Security Act of 2018 (FASCSA). Through letters dated March 3, 2026, the department formally notified Anthropic of its designation, the first such action ever applied to an American company. The breakdown in negotiations centered on Anthropic's refusal to grant the department unfettered access to its Claude models for all lawful purposes; Anthropic sought assurances that the technology would not be used for fully autonomous weapons or domestic mass surveillance. The designation requires defense vendors and contractors to certify that they do not use Anthropic's models in their Pentagon work.
**Key litigation posture:**
– On March 9, Anthropic filed two lawsuits in federal court challenging the designation on statutory and constitutional grounds.
– U.S. District Judge Rita Lin issued a ruling blocking enforcement of the administration's directive ordering all federal agencies to stop using Anthropic products.
– On April 8, the D.C. Circuit Court of Appeals denied Anthropic's request to pause the supply chain risk designation pending appeal, with oral argument set for May 19.
– Government agencies including the Pentagon remain able to use Anthropic's models while the legal fight continues; the National Security Agency is reportedly using the company's Mythos model.
The Pentagon's announcement runs in parallel with a separate White House effort to canvass the private sector on AI security threats. Officials from the White House's Office of the National Cyber Director reached out to tech companies with questions covering AI, information sharing, vulnerability patching, and the appropriate federal role. That outreach followed an April 28 meeting between White House cybersecurity officials and tech industry representatives focused on dangers and opportunities posed by powerful new AI models. Industry reception was uneven: POLITICO reported that some representatives were uncertain about the purpose of the questions, while others declined to disclose security protocols without a clearer justification. [POLITICO]
The May 1 contracts extend a procurement trajectory that accelerated sharply once the Anthropic dispute crystallized. Anthropic had signed a $200 million contract with the department in July 2025 and was the first AI lab to integrate its models into mission workflows on classified networks. Friday's agreements build on Secretary Pete Hegseth's push to bring commercial AI into the department, including the December launch of the GenAI.mil platform. That platform has been used by more than 1.3 million Defense Department personnel in its first five months of operation. Pentagon Chief Technology Officer Emil Michael, commenting on the new agreements, framed the multi-vendor approach as an explicit lesson from the Anthropic episode, emphasizing the need for diversity of supply across both open-source and proprietary models. [4]
Diplomatic signals suggest the Anthropic dispute may not be permanent. The White House is developing guidance that would allow agencies to work around the supply chain risk designation and onboard new Anthropic models, including Mythos, according to sources familiar with the matter. A draft guidance document would permit agencies to sidestep the designation without formally rescinding it, creating a possible offramp while leaving the Pentagon's underlying position unresolved. President Trump separately signaled a warmer tone last week, saying in a CNBC interview that Anthropic is "shaping up" and could "be of great use" in the future. Whether any accommodation extends to classified network access, as opposed to civilian agency use, remains the operative legal and policy question.
—
—
**References:**
[1] Reuters. (2026, May 1). Pentagon reaches agreements with leading AI companies. https://www.investing.com/news/stock-market-news/pentagon-reaches-agreements-with-leading-ai-companies-4652828 [2] NBC News. (2026, April 29). Pentagon inks deal with Google for AI services. https://www.nbcnews.com/tech/tech-news/pentagon-inks-deal-google-ai-services-rcna342661 [3] The Hill. (2026, May 1). Pentagon reaches deal with leading AI companies for classified work. https://thehill.com/policy/technology/5858995-pentagon-ai-companies-classified-work-deal/ [4] Breaking Defense. (2026, May 1). Pentagon clears 8 tech firms to deploy their AI on its classified networks. https://breakingdefense.com/2026/05/pentagon-clears-7-tech-firms-to-deploy-their-ai-on-its-classified-networks/ [5] Department of War. (2026, May 1). Classified Networks AI Agreements. https://www.war.gov/News/Releases/Release/Article/4475177/classified-networks-ai-agreements/ [6] CNBC. (2026, March 5). Anthropic officially told by DOD that it's a supply chain risk. https://www.cnbc.com/2026/03/05/anthropic-pentagon-ai-claude-iran.html [7] Mayer Brown. (2026, March 27). Anthropic Supply Chain Risk Designation Takes Effect — Latest Developments and Next Steps for Government Contractors. https://www.mayerbrown.com/en/insights/publications/2026/03/anthropic-supply-chain-risk-designation-takes-effect–latest-developments-and-next-steps-for-government-contractors [8] NPR. (2026, March 6). Pentagon labels AI company Anthropic a supply chain risk 'effective immediately.' https://www.npr.org/2026/03/06/g-s1-112713/pentagon-labels-ai-company-anthropic-a-supply-chain-risk [9] PBS NewsHour / AP. (2026, April). White House chief of staff to meet with Anthropic CEO over its new Mythos AI model. https://www.pbs.org/newshour/politics/white-house-chief-of-staff-to-meet-with-anthropic-ceo-over-its-new-mythos-ai-model [10] Axios. (2026, April 29). Scoop: White House workshops plan to bring back Anthropic. https://www.axios.com/2026/04/29/trump-anthropic-pentagon-ai-executive-order-gov [11] Cybersecurity Dive. (2026, April 30). White House questions tech industry on defensive AI use, cybersecurity resilience. https://www.cybersecuritydive.com/news/white-house-oncd-ai-tech-industry-questions/819133/ [12] Nextgov/FCW. (2026, April 29). White House is drafting plans to permit federal Anthropic use. https://www.nextgov.com/artificial-intelligence/2026/04/white-house-drafting-plans-permit-federal-anthropic-use/413202/
Comments (0)