Pentagon inks deals with Nvidia, Microsoft and AWS to deploy AI on classified networks

After landing agreements with Google, SpaceX, and OpenAI, the U.S. Defense Department said connected Friday that it has signed deals with Nvidia, Microsoft, Amazon Web Services, and Reflection AI that let it to deploy their AI tech and models connected its classified networks for “lawful operational use.”

“These agreements accelerate the translation toward establishing the United States subject arsenic an AI-first warring unit and volition fortify our warfighters’ quality to support determination superiority crossed each domains of warfare,” the connection reads.

The deals travel arsenic the U.S. Department of Defense has accelerated its diversification of AI vendors successful the aftermath of its arguable dispute with Anthropic implicit usage presumption of its AI models. The Pentagon wanted unrestricted usage of Anthropic’s AI tools, but the AI laboratory insisted connected guardrails to forestall Anthropic’s tech from being utilized for home wide surveillance and autonomous weapons.

The 2 are warring it retired successful tribunal astatine the moment, though Anthropic successful March won an injunction against the Pentagon’s determination to marque the institution a “supply concatenation risk.”

“The Department volition proceed to physique an architecture that prevents AI vendor lock-in and ensures semipermanent flexibility for the Joint Force,” the connection reads. “Access to a divers suite of AI capabilities from crossed the resilient American exertion stack volition springiness warfighters the tools they request to enactment with assurance and safeguard the federation against immoderate threat.”

The DoD said the companies’ AI hardware and models volition beryllium deployed connected Impact Level 6 (IL6) and Impact Level 7 (IL7) environments to “streamline information synthesis, elevate situational understanding, and augment warfighter decision-making.” IL6 and IL7 are high-level information classifications for information and accusation systems that are deemed captious to nationalist security, and necessitate that these systems beryllium protected physically, done strict entree controls and audits.

The Pentagon said much than 1.3 cardinal DoD unit person truthful acold utilized its unafraid endeavor level for generative AI, GenAI.mil, which provides entree to ample connection models (LLMs) and different AI tools wrong government-approved unreality environments. It is designed to assistance chiefly with non-classified tasks similar research, papers drafting, and information analysis.