OpenAI wants to rewrite its Washington playbook with reverse federalism strategy

At a glance:

• OpenAI is pursuing a state-by-state lobbying campaign it calls "reverse federalism" to shape AI regulation after the Trump administration pulled back on federal AI oversight.
• The effort has already targeted California and New York, and is now pushing an AI safety bill in Illinois that would require frontier labs to submit to audits.
• Anthropic is countering in Illinois with its own competing bill, setting up a high-stakes policy fight in one of the most populous states.

What happened with the federal executive order

The Trump administration was poised to sign an executive order on Thursday that would have created a voluntary framework for frontier AI companies to have their models reviewed by the federal government for security risks before going public. But the order was reportedly canceled — not because of any substantive policy disagreement, but because tech executives couldn't get to Washington fast enough for the photo opportunity. Donald Trump later claimed he "didn't like certain aspects of it" and doesn't want to "do anything that will get in the way" of America's AI competition with China, according to reporting from PunchBowl News.

The incident underscores a broader pattern of disinterest from the current administration toward regulating AI companies. The administration has shown a willingness to penalize firms that draw its ire — swiftly and immediately — but otherwise has taken a hands-off approach to AI governance. This posture has persisted even as the administration tried to force that same hands-off approach onto states by pushing a 10-year moratorium on any AI regulation, an effort that has repeatedly failed.

OpenAI's "reverse federalism" play

With federal regulation unlikely and state policy up for grabs, OpenAI has moved aggressively to shape the regulatory landscape from the ground up. According to a Politico report, the company has launched a state-by-state lobbying effort designed to make the patchwork of state-level AI rules as favorable as possible across the country. OpenAI's chief global affairs officer, Chris Lehane, calls the plan "reverse federalism" — the idea being that by passing similar laws in enough states, the company can create something that functions like a unified national regulatory framework without waiting for Congress or federal agencies to act.

The strategy has already delivered results in two of the most influential states. California passed an AI safety law late last year that OpenAI ultimately endorsed, and New York likewise passed its own AI safety law. Now the company is turning its attention to Illinois, where it has offered its endorsement for an AI safety bill that would require frontier labs to submit to audits.

Notably, the bill OpenAI backed shares a lot of similarities with the laws already passed in New York and California, suggesting the company is templating its regulatory approach across states to build consistency. Should the Illinois bill pass, OpenAI would have successfully shaped the foundational policy it is subject to in three of the most populous and influential blue states in the country.

Anthropic fights back in Illinois

OpenAI does have one major hurdle to completing the trifecta: Anthropic. The rival AI lab is also reportedly working behind the scenes in Illinois to guide policy and has thrown its weight behind a competing bill. That puts the two leading frontier AI labs on opposing sides of a single state's legislative process, raising the stakes for whichever bill ultimately moves forward.

The Illinois fight is a microcosm of the broader battle playing out across the country. With no federal backstop and a patchwork of state rules in flux, AI companies are investing heavily to ensure that the rules they end up operating under are written — or at least heavily influenced — by themselves.

The money behind the lobbying

The state-by-state campaign is not cheap. CalMatters estimated that tech companies broadly spent about $40 million last year on lobbying efforts in California alone, and efforts have only ramped up this year. Millions are flowing into pro-AI super PACs designed to influence public opinion on AI policy, adding a campaign-style dimension to what has traditionally been a back-room regulatory game.

As these AI giants continue to raise more and more money at what many consider absurd valuations, larger portions of those funds are being directed toward shaping the laws that will govern their businesses. The spending pattern mirrors what happened in earlier technology cycles — telecommunications, social media, autonomous vehicles — but the pace and intensity appear to be accelerating as AI becomes a national security and economic priority.

Why Trump's non-governance strategy matters

In a sense, Trump's strategy of not governing has worked as intended — at least from the perspective of the companies filling the vacuum. With the federal government stepping back from AI regulation, the field has been left open for AI companies to dictate their own rules through state-level action, lobbying, and political spending. The result is a system where the regulated are also the regulators, or at least the ones writing the first drafts.

This dynamic raises questions about accountability and public input. Traditional regulatory processes usually involve public comment periods, hearings, and agency expertise. When companies are effectively writing the rules through state legislatures and PACs, the role of government shifts from referee to spectator. For investors, employees, and the public, the implications of that shift are only beginning to become clear.

What to watch next

The Illinois bill will be a key data point. If OpenAI's endorsed version passes, it could embolden the company's "reverse federalism" playbook in other states, potentially creating a de facto national standard built state by state. If Anthropic's competing bill wins, it could signal that the strategy is running into resistance from well-funded rivals.

At the federal level, the canceled executive order suggests the Trump administration remains unlikely to impose top-down AI rules, leaving the state-by-state race as the primary battleground for the foreseeable future. Investors and policymakers should monitor lobbying spend, bill language across states, and the positions of other frontier labs as the regulatory landscape continues to take shape without federal guidance.