US Gov inks blanket deal for Google, OpenAI, Anthropic AI products
The United States government is significantly accelerating its adoption of artificial intelligence, with the General Services Administration (GSA) announcing a blanket agreement that streamlines the procurement of AI products from leading firms Anthropic, Google, and OpenAI. This move is designed to make it far easier for federal, state, and local agencies to integrate advanced AI capabilities into their operations.
Under the new arrangement, Anthropic’s Claude, Google’s Gemini, and OpenAI’s ChatGPT have been added to the GSA’s Multiple Award Schedules (MAS). These MAS contracts are a crucial mechanism, allowing companies to sell their products to government entities over extended periods without the need for individual contract negotiations with each agency. GSA leadership has framed this initiative as a direct implementation of President Trump’s recently unveiled AI action plan, signaling a concerted effort to embed AI across the federal landscape.
Josh Gruenbaum, Commissioner of the Federal Acquisition Service, emphasized the broad spectrum of solutions now available. Agencies, he noted, will be able to explore everything from foundational large language models for research assistance to highly specialized, mission-critical applications. This strategic shift is expected to support a diverse range of functions, from automating back-office tasks to enabling real-time language translation, bolstering cybersecurity defenses, and facilitating large-scale data analysis. The GSA itself has already reported significant efficiency gains from early AI adoption, saving an estimated 365,000 staff hours in 2025 alone.
A notable component of this new agreement is a specific partnership with OpenAI. Under this deal, federal agencies will gain access to ChatGPT’s Enterprise tools for a nominal annual fee of just one dollar, coupled with comprehensive training for federal employees provided by OpenAI. While similar explicit announcements regarding Google and Anthropic were not made, the general MAS inclusion still provides a clear path for agencies to acquire their offerings.
Looking ahead, the GSA is committed to a cautious, security-first approach to AI procurement, ensuring sensitive information remains protected while still enabling efficiency gains. This includes a focus on models that prioritize truthfulness, accuracy, transparency, and freedom from ideological bias, aligning with the Trump administration’s policy that federally procured AI systems must adhere to these principles. While the current roster includes the tech giants, the GSA is also exploring additional partners. Clues from a previously leaked GitHub page outlining the administration’s “whole-of-government AI plans” suggest that other vendors, including Meta’s Llama and AI firm Cohere, are under consideration, though Cohere’s lack of FedRAMP certification for secure cloud computing was noted.
Despite the streamlined purchasing process, significant hurdles persist. The Government Accountability Office (GAO) has highlighted ongoing concerns among many agencies regarding AI reliability, the potential for biased or incorrect outputs, and a general lack of model transparency. These fundamental issues, auditors point out, are not inherently resolved by making the procurement process simpler. The Trump administration has consistently advocated for widespread AI integration to reduce spending and optimize operations, yet previous attempts at deployment have often stalled due to funding constraints and excessive regulation. While these new MAS additions might spur purchasing, they do not automatically mitigate the deeper, systemic challenges associated with deploying artificial intelligence responsibly and effectively across a vast government apparatus.