Meta's Llama System Gets Greenlit For US Government Use In Big AI Push
The approval, a key component of the Trump administration's broader initiative to integrate commercial AI tools, adds Meta’s Llama models to the GSA’s list of approved technologies under the OneGov program. The purpose of this initiative is to streamline government procurement by providing a centralized list of vetted, secure, and legally-compliant tools, thus eliminating duplicate work between agencies.
According to a press release by Meta, the collaboration will not just make the open-source and free Llama models more accessible to federal agencies, but because of Llama's open-source nature, government departments can maintain full control over their data. Llama's flexibility should also allow agencies to build, deploy, and scale AI applications at a lower cost and in turn lower the burden on American taxpayers. Meta CEO Mark Zuckerberg commented that Meta wants to ensure that “all Americans see the benefit of AI through better, more efficient public services.”
Josh Gruenbaum, the GSA's procurement lead, emphasized that the goal of integrating such tools is to improve the country, not to benefit any single company or curry favor with the president. For now, Llama is expected to assist federal agencies with a variety of tasks, from speeding up contract reviews to resolving IT issues more efficiently. This partnership is seen as strengthening the government's capacity to test and deploy AI while safeguarding sensitive information, aligning with the objectives of America's AI Action Plan.
With this deal, Meta joins a trending number of big tech companies offering AI products to the government, often at a reduced cost or for free (as with the Llama deal) while ensuring they meet federal security standards. For example, OpenAI and Anthropic are offering models at a nominal $1 per agency per year, while Google’s Gemini for Government is priced at $0.47 per agency annually. Microsoft has also secured a deal to provide its products, including Copilot AI, to the GSA at a discounted rate.
Meanwhile, the surge in these deals has prompted the Federal Risk and Authorization Management Program (FedRAMP) to prioritize the authorization of certain AI cloud services.
