Artificial intelligence continues to evolve and outpace even the most attentive observer's ability to keep up, and federal AI policy and actions are no exception. It is essential contractors stay engaged with the shifting federal AI landscape.

While the federal AI space has not seen sweeping new mandates or policy updates recently, activity has not been completely quiet. A handful of agency announcements and emerging regulatory mentions suggest AI continues to take shape at federal agencies. The following selected news roundup helps to keep readers informed on recent federal AI actions and what may be expected on the horizon.  

DOD/W Classified Networks AI Agreements

At the beginning of May 2026, the Department of Defense/War (DOD/W) announced agreements with eight frontier AI companies: SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, Amazon Web Services, and Oracle. The agreement stipulates that these companies can deploy advanced AI capabilities on the DOD/W classified networks under lawful operational purposes. The goal of the integration into classified settings is to, “streamline data synthesis, elevate situational understanding, and augment warfighter decision-making in complex operational environments,” according to the announcement. The agreement also comments on the expansion of the GenAI.mil environment, signaling the department’s operationalization of AI at the enterprise level.

For contractors, the agreement with these eight vendors shifts the landscape of DOD/W’s preferred AI technology stack for classified environments. Nonetheless, the multi-vendor AI approach prevents vendor lock-in and allows opportunities for systems integrators to bridge the different AI ecosystems.

GSA to Charge for USAi Platform

Last year, GSA launched USAi.gov, a free and secure evaluation site for federal agencies to experiment with generative AI solutions. Agencies can use the platform to test AI solutions against mission delivery, improving customer experience, analyzing data, automating workflows, and more. According to GSA FY 2027 budget documents, “the current product is being used by 15 pilot agencies with a substantial list of agencies in the waiting list.” Given the platform’s popularity, GSA budget documents state that USAi.gov will be offered under reimbursable models beginning in FY 2027.

FY 2026 NITRD AI PCA

Last week, the Networking and Information Technology Research and Development (NITRD) released its annual supplement to the president’s budget request. Though more than halfway through the FY 2026 cycle, this recent supplement only covers the FY 2026 time frame. The purpose of the supplement is to provide the research and development budgets of participating NITRD agencies, which include areas such as artificial intelligence, networking, advanced wireless communications, and more.

The AI Program Component Area (PCA) covers R&D programs such as strengthening human-AI collaboration to enhance U.S. productivity; security and resilience of AI systems; building and expanding access to high-quality datasets, testbeds, and AI training resources, among others. The supplement cites a total of $1.8B in AI R&D in FY 2026, a $523M decrease from the $2.3B FY 2025 estimated level. The supplement also identifies $988M in additional agency cross-cutting AI-relevant R&D investments by other NITRD PCAs.

While the supplement stops short of providing a concrete explanation specific to the AI PCA's budget decline, it does offer a broader rationale for reductions across most PCAs in FY 2026. The document cites two factors for the decreases: first, actual and estimated investments in high-priority IT R&D areas have historically outpaced requested amounts, as more proposals are submitted and funded than originally anticipated. Second, agencies are actively pursuing efficiency improvements, cost savings, and government modernization initiatives, which collectively result in budget reductions.

AI Requirements in GSA Refresh 32

Previously, I wrote on an AI clause that GSA was initially considering for MAS Refresh 31, in which the agency ultimately decided to consider the clause for MAS Refresh 32 in response to a request for an extension by industry. The clause depicts government data rights and safeguards within GSA Schedule solicitations and contracts for AI solutions.

According to an Advanced Notice for MAS Refresh 32 from GSA on the upcoming refresh in June 2026, the announcement does not list the incorporation of an AI clause in the next refresh. Instead, the announcement lists implementation of Executive Order 14398 and incorporation of FAR 52.222-90. Though the AI clause may still be on the table for the next refresh since GSA does not specifically state that it has been postponed once again, industry should keep a close eye on the draft attachment expected by GSA prior to issuance of MAS Refresh 32.

Potential AI Security Executive Order

According to a Federal News Network article, the White House is considering the release of a new AI security executive order. “We’re studying possibly an executive order to give a clear road map to everybody about how this is going to go and how future AI that also potentially create vulnerabilities should go through a process so that they’re released in the wild after they’ve been proven safe, just like an FDA drug,” according to Kevin Hassett, director of the National Economic Council. Should such an executive order be released, contractors will have a formal pre-deployment evaluation requirement within federal AI acquisitions.

Reinstating Federal Use of Anthropic

According to an April 2026 Government Executive article, the Trump Administration is reconsidering its federal prohibition on Anthropic AI products and services. As mentioned previously, the White House and federal agencies banned the use of Anthropic AI tools due to a dispute on the boundaries of acceptable use between the contractor and DOD/W, and even designated Anthropic as a supply chain risk.

Since then, Anthropic announced its Mythos AI model, to be used in advanced cyber operations for vulnerability detection. As a result, the new model is catching the attention of many government leaders. According to the article, “President Donald Trump said in a CNBC interview that the company is “shaping up” and can “be of great use” in the future, a sign that tensions between Anthropic and the government may be easing.” For contractors, reinstatement would reopen the competitive landscape a bit more again. Moreover, agencies with cyber-focused missions such as NSA, CYBERCOM and CISA, would likely be early adopters of Mythos, potentially creating a downstream effect of additional and related opportunities for contractors.

National AI Legislation by 2027

The Office of Science and Technology Policy Director, Michael Kratsios, has been vocal on the push for a detailed AI legislation by the end of 2026, as called for by Executive Order (EO) 14365, Ensuring a National Policy Framework for Artificial Intelligence. “We’re in the early innings, but it’s the important sort of catalyst for what will be a big part of the agenda this year,” said Kratsios, per an article by GovCIO Media & Research. The framework is envisioned as a unified federal approach to AI governance vs. the fragmented nature of state AI regulations. As contractors await Congress and the White House to work towards the effort, it will be interesting to see the balance struck between governance and innovation in the upcoming legislation.