GSA introduces USAi.Gov to streamline AI adoption across government

The General Services Administration unveiled another major Trump administration initiative Thursday, launching USAi.Gov to serve as a centralized, secure and trustworthy generative artificial intelligence evaluation suite for federal agencies to explore, interact with and even adopt secure AI services.

GSA’s goal is to streamline government adoption of the emerging technology by offering AI products like chatbots, code generation, document summarization, large language models and other generative AI tools to agencies in a trusted environment. 

The launch supports priorities called for in President Donald Trump’s AI Action Plan issued in July regarding the creation of an effort to accelerate AI adoption across government and develop an AI evaluation suite. 

“USAi means more than access—it’s about delivering a competitive advantage to the American people,” GSA Deputy Administrator Stephen Ehikian said in a statement. “The launch of USAi shows how GSA is translating President Trump’s AI strategy into action and accelerating AI adoption across government. USAi will put mission-ready tools directly into the hands of agencies to modernize faster, boost security, and lead globally.”

At launch, USAi will include AI models from Anthropic, Google, Meta and OpenAI. GSA had considered including models from up to six other firms at launch, including AI21 Labs, Amazon, Cohere, Mistral AI, Stability AI and xAI, according to early demos viewed by Nextgov/FCW.

According to GSA, agencies will be able to “explore capabilities, measure performance and identify strengths and limitations across different AI systems,” empowering agencies to make informed decisions about the adoption, customization and integration of myriad AI tools. Agencies can use the shared services platform to efficiently test and evaluate emerging technologies in a secure, cloud-based environment that supports scalable solutions.

In other words, agencies now have a place to try these technologies before they buy them. 

“USAi isn’t just another tool, it’s infrastructure for America’s AI future,” said GSA Chief Information Officer David Shive. “USAi helps the government cut costs, improve efficiency, and deliver better services to the public, while maintaining the trust and security the American people expect.” 

Many of USAi.Gov’s relevant features adhere to key policy and security requirements issued by the Trump administration. Early demos of the platform featured centralized logging, bias evaluation and inventory support in line with the 2025 Office of Management and Budget AI Governance Memorandum, as well as guardrail enforcement compliant with the 2025 Executive Order on Trustworthy AI issued by Trump in June. The government-wide effort is also entirely compliant with evolving cloud security standards through the Federal Risk and Authorization Management Program, or FedRAMP, which was first launched in 2011.

Per GSA, USAi.Gov “further advances workforce readiness by providing dashboards and usage analytics that help agencies track performance, measure maturity, and guide adoption strategies. By offering a centralized environment for experimentation, the platform supports both workforce upskilling and agency-level digital transformation.”

“GSA is proud to be on the frontlines of delivering on President Trump’s call to accelerate AI adoption in government and help empower agencies to deliver enhanced services to the American public," said FAS Commissioner Josh Gruenbaum. "USAi provides a secure and collaborative environment for government employees to explore cutting-edge AI models that will make day-to-day workflows more efficient and help discover innovative solutions prior to making procurement decisions."

More momentum at GSA

USAi.Gov is the latest in a series of seismic efforts undertaken by GSA. 

The agency has already secured multiple deals with major companies, including Google Public Sector, Adobe, Salesforce, Elastic, Oracle, Uber, IBM, Docusign, Amazon Web Services, OpenAI, Anthropic and Box. The resulting deals have ultimately discounted billions of dollars’ worth of software those companies provide government customers through the OneGov initiative, which essentially treats government customers as a single large purchaser. Much of the tech developed by those firms includes AI, cloud computing and emerging capabilities necessary for the administration of government in the modern age.

GSA has also undertaken an aggressive rewrite of Federal Acquisition Regulation rules and centralized approximately $400 billion worth of goods and services through the agency, using Trump’s April 15 executive action as marching orders. The agency is also considering moving the NASA SEWP and NIH contract vehicles under its management to further streamline acquisition operations. It’s also taken on an expansive review of dozens of consultancies and has doubled its number of FedRAMP authorizations over last year’s numbers, part of a major FedRAMP overhaul. 

Per Ehikian, GSA itself is a regular consumer of AI, with more than half of its staff using internal AI tools daily.  

The agency has accomplished this despite reductions in force, more than 2,100 GSA staff accepting deferred resignation offers and recent leadership changes. On Aug. 1, Trump nominated financial executive Edward Forst to be GSA’s administrator; Michael Rigas is currently serving as the agency’s acting administrator. 

AI is all the rage in government

The federal government has generally increased the amount it spends each year on technology across civilian agencies and the Defense Department since 2015. 

But studies indicate its investment in AI has rapidly outpaced traditional IT investments. A 2024 Brookings report examining federal contracting data found massive growth in AI spending from August 2022 to August 2023, with obligated funds increasing 150% in AI expenditures over various federal agencies. The potential value of contract awards during the same time increased an exponential 1,200%, from $355 million to $4.5 billion.

“Growth in private sector R&D has been at above 10% per year for a decade while the federal government has shown more modest growth over the last five years after a period of stagnation,” the report states. “However, the 1,200% one-year increase in AI potential value of awards of over $4.2 billion is indicative of a new imperative in government AI R&D leading to deployment.”

Even so, the federal government has had a difficult time tracking all its AI efforts. In a landmark Government Accountability Office audit in 2023, the watchdog found some 1,200 planned or operational AI use cases across agencies. Yet the report found the government, just like the private sector, struggles in even defining what AI is, and those complications are furthered by the mainstream introduction of large language models and generative AI tools.

“AI capabilities are rapidly evolving, but neither the scientific community nor industry agree on a common definition for these technologies. Even within the government, definitions vary,” the report said.

USAi.gov may address that problem.

”Because AI is both powerful and transformative and also challenging if not dangerous, the more transparency around it, the better,” said Stan Soloway, president and CEO of Celero Strategies and former defense acquisition official. “Trying to cleanly and quickly find how and where AI capabilities are being used across the government is very difficult. If GSA can help solve some of those problems, that is all to the good.” 

USAi.Gov’s launch comes one month after the Defense Department issued contracts worth up to $200 million each to four companies: Anthropic, OpenAI, Google and xAI. Three of the companies’products and models will be found on the USAi platform, and in the past week,  Anthropic and OpenAI agreed to provide their software to government agencies for nominal fees. 

xAI, founded by billionaire and former Department of Government Efficiency leader Elon Musk, recently unveiled its new Grok for Government offering in a post on X. Very soon, the featured chatbot on the X social media platform could be put to use by federal employees to deliver citizen services.

The recent maneuvering by tech firms is an indication of how important AI and emerging technologies have become to federal agencies under Trump, who has vowed to keep the U.S. government ahead in the global AI race. 

About those companies

Anthropic and OpenAI, well-known for their large language models, are relatively new to the government market but have both made major splashes. Direct competitors, both companies just won multi-million dollar defense contracts, are building out teams to hone in on public sector business and have large language models that are widely used by the public. Anthropic’s cloud partner of choice is Amazon Web Services; OpenAI’s primary cloud partner is Microsoft, but the company recently announced a partnership with competitor Google to beef up cloud computing capacity.

Meta, founded by billionaire Mark Zuckerberg, is no stranger to the public, due to its social media platform, but it is relatively new to serving government customers. In late 2024, the New York Times reported Meta began allowing federal agencies and select defense contractors, including Booz Allen Hamilton, Lockheed Martin and Palantir, to use its Llama AI models. It has since partnered with Anduril to develop the Army’s multibillion dollar Soldier-Borne Mission Command Next program, formerly called IVAS Next.

Google, meanwhile, is one of the largest tech firms in the world, and has been serving the U.S. government for many years, especially through its hyper-scale cloud computing business and Workspace suite. Google’s Vertex AI is among the most widely used platforms for building and deploying machine learning and other AI models on. In April, Google became the first tech firm to offer a significant government discount of its software, which the company estimated could save federal agencies as much as $2 billion over the next three years.