OpenAI opens advanced AI models to all vetted government agencies

OpenAI announced that its latest generation of AI models will be available to every vetted level of U.S. government—from federal agencies down to state and local offices—giving verified users special versions with reduced safety guardrails to help shore up cyber defenses [1].

Democratizing cyber‑defense tools

OpenAI’s “Trusted Access for Cyber” program, previously limited to a handful of partners, is being broadened to any government entity that passes verification. Sasha Baker, head of national security policy, said the move is meant to “democratize our ability to uplift everyone who needs cyber defense” rather than reserving the technology for the Fortune 50 [1]. The company framed the decision as a response to a “wake‑up call” in the cybersecurity community, arguing that broader access can pre‑empt malicious exploitation of AI‑driven vulnerabilities.

Stakes for OpenAI’s broader business

The expansion comes as OpenAI grapples with internal growth targets. The Wall Street Journal reported that the firm fell short of its goal of one billion weekly ChatGPT users by the end of 2025, reaching 900 million in February 2026—a 125 % year‑over‑year increase but still below the internal benchmark [2]. At the same time, OpenAI’s annualised revenue sits at roughly $25 billion, while the company has committed to $600 billion in compute spending through 2030 to sustain its model training pipeline [2]. The new government access program is positioned as a way to generate additional demand for its AI services, potentially offsetting the revenue gap that analysts fear could jeopardize its massive infrastructure contracts with firms like Oracle and CoreWeave.

Competitive context

OpenAI’s approach contrasts sharply with Anthropic’s more cautious rollout of its Mythos model, which is limited to a tightly controlled consortium despite similar cybersecurity capabilities [1]. Anthropic argues that tighter controls are needed to curb an AI‑driven arms race, while OpenAI is betting on broader distribution to accelerate defensive capabilities. The divergent strategies highlight a broader industry split: firms must balance rapid innovation against the risk of enabling malicious actors.

What to watch

• Policy guidance – OpenAI said it will take direction from the White House on how to shape its AI‑government partnership; upcoming guidance could define the scope of future access.
• Rollout milestones – The company plans a follow‑up visit to Washington in the coming weeks to gather feedback; the speed and depth of implementation will signal how quickly agencies adopt the tools.
• Revenue impact – Monitor OpenAI’s quarterly earnings for any uptick in government‑related contracts that could help bridge the gap between current $25 billion revenue and the $600 billion compute commitments.

By opening its most capable models to every vetted government tier, OpenAI is testing whether a broader defensive ecosystem can sustain its ambitious growth assumptions, while the industry watches to see if the strategy narrows the gap between AI capability and cybersecurity risk.