Pentagon inks AI procurement deals with seven companies, leaves out Anthropic

The U.S. Defense Department today announced that it has inked artificial intelligence procurement contracts with seven tech firms.

The group includes Amazon Web Service Inc., Google LLC, Microsoft Corp., Nvidia Corp., OpenAI Group PBC and SpaceX Corp. They are joined by a startup called Reflection AI Inc. that raised $2 billion in funding last year. Notably absent from the list is Anthropic PBC, which the Trump administration designated as a supply chain risk in February.

The Pentagon will make the participating companies’ products available through an internal AI portal called GenAI.mil. According to GeekWire, more than 1.3 million Defense Department personnel have adopted the platform since its launch last year. Those users have built hundreds of thousands of AI agents.

The AI products covered under the newly issued contracts will be used in Impact Level 6 and Impact Level 7 environments. Those are Defense Department systems that can be used to store classified information. According to Nextgov/FWC, the goal of the AI initiative is to ease data synthesis and streamline decision-making workflows.

It’s unclear what AI products the contracts encompass. Nvidia, which is best known for its graphic cards, also offers neural network development tools. Additionally, it has released an open-source large language model series based on a Mamba-Transformer architecture. The Transformer architecture underpins most LLMs, while Mamba is a competing technology with lower memory requirements.

SpaceX, which has also won an AI contract from the Pentagon, became an LLM provider earlier this year after it merged with xAI Holdings Corp. The deal bought the aerospace company the Grok family of language models. SpaceX may soon expand its AI portfolio by acquiring Cursor’s Composer series of coding models.

Reflection AI, the only company on today’s list of Pentagon contract winners that isn’t a household name, was founded in 2024 by former Google DeepMind researchers. It hasn’t yet announced a commercial product. Last year, TechCrunch reported that Reflection AI plans to release a language model trained on tens of trillions of tokens.

The Pentagon’s decision not to issue a contract to Anthropic isn’t surprising. In March, Defense Secretary Pete Hegseth designated the company a supply chain risk. The designation prohibits the Pentagon from using Claude and also limits defense contractors’ access to the LLM series. 

Hegseth announced the decision after Anthropic declined to let the Pentagon use Claude for “all lawful” purposes. The company was concerned that such contractual language would allow its AI models to be used for domestic mass surveillance or the development of autonomous weapons. In March, Anthropic filed suit to challenge its designation as a supply chain risk.

The Defense Department is reportedly using the company’s Claude Mythos Preview model despite the ban. A number of other federal agencies have also adopted the model. Anthropic announced Claude Mythos Preview last month, but has not yet made the LLM publicly available because it’s highly adept at finding zero-day or yet-undiscovered cybersecurity vulnerabilities. 

Photo: David B. Gleason/Flickr