Dean Ball helped write the federal government’s AI rulebook. In July, he starts work at OpenAI. That one sentence is the AI policy revolving door in miniature, and the door is spinning in a clear direction.

Ball served as senior policy adviser for AI and emerging technology at the White House Office of Science and Technology Policy, where he was a primary author of the administration’s AI Action Plan, the document that set federal expectations on chip exports, AI safety, and how Washington deals with the companies building frontier models. On July 6 he becomes head of a new OpenAI team called Strategic Futures, reporting to chief strategy officer Jason Kwon and working on catastrophic risk, recursive self-improvement, labor-market effects, and the relationship between frontier labs and governments. He keeps a non-resident fellowship at the Foundation for American Innovation. The same week, Noam Shazeer, who co-wrote the 2017 paper that made modern AI possible, left Google for OpenAI. One of those moves is talent. The other is governance.

OpenAI did not respond to a request for comment.

Here is the part worth sitting with: the AI policy revolving door is not new. The door between Washington and industry has been turning for decades, and it turns in both parties. During the Obama years, the Tech Transparency Project counted 258 revolving-door moves between Google and the federal government. The Pentagon version is older and larger, with hundreds of senior defense officials cycling into contractor boardrooms, a pattern the Project On Government Oversight has tracked for years. When Biden staffed his administration, advisers openly described technology firms as the new Goldman Sachs, the way that bank once seeded every Treasury. We covered the policy-text version of this in how the AI executive order got written, and the thinning line between state and company in Britain’s sovereign AI push. Ball is the AI era’s turn of a very old wheel.

That history is also Ball’s defense, and it is a fair one. Government needs people who actually understand the technology, and you do not get that understanding without moving talent in and out of the field. The door has always swung both ways. Someone who helped write a framework is not a regulator signing off on OpenAI’s compliance, and keeping a public fellowship is more transparency than most bother with. The distinctions are real.

The mistake is hunting for the villain. There isn’t one, and that is the whole problem. Power has stopped needing corruption now that it can simply hire the referee. The defense industry took generations to perfect this move, Wall Street took decades, and AI ran the same play in about three years, in the open, announced over press releases. What we are watching is a narrow class of people learning to write the rules and own the upside inside the same career, and calling the combination expertise. The public was never at that table. It only gets the bill.

For everyone outside this world, the takeaway is plain. The AI policy revolving door is why the rules about the AI in your bank, your hospital, and your benefits keep getting written by a circle of people who end up at the companies those rules cover. No one has to break anything for the public to lose its seat at the table, and that is a harder problem to fix than corruption, because nothing illegal ever happens.

Watch who moves next. The names leaving government for the labs are a better map of where AI policy is actually heading than anything published in the Federal Register.

FAQ

Who is Dean Ball?
Dean Ball was the senior policy adviser for AI and emerging technology at the White House Office of Science and Technology Policy and a primary author of the administration’s AI Action Plan. In July 2026 he joins OpenAI to lead a new team called Strategic Futures.

What is the AI policy revolving door?
The AI policy revolving door is the movement of people between the government roles that write AI rules and the companies those rules govern. The pattern is not unique to AI. It has long run between the Pentagon and defense contractors, and between the Treasury and Wall Street.

Related Stories

Lamar Rice
Editor, Laterstack

Electricity bills across the United States are climbing and people in several states are blaming one thing. The nonstop growth of data centers that power modern artificial intelligence systems. Residents in Virginia, Illinois and Ohio saw double digit increases in utility costs this year. In each of these states there is a dense cluster of massive AI facilities running around the clock.

Energy regulators say the math is simple. A single large scale data center can draw as much power as hundreds of thousands of homes. When dozens of them appear in the same region they reshape the entire grid. When demand shoots up faster than new power plants can be approved or built, prices follow.

Virginia provides the clearest example. The state has the highest concentration of data centers anywhere in the world. Local leaders have begun openly targeting the industry. Newly elected Governor Abigail Spanberger rode a campaign focused on affordability and promised voters that tech companies would pay more of the costs created by their facilities.

A political fight that is growing louder

The energy crunch arrives at a sensitive moment. National elections sit just one year away and electricity bills have become a daily conversation. Several Democrats in Washington now argue that the relationship between President Trump and major AI companies has allowed utilities to pass data center costs to ordinary families.

Senators Bernie Sanders and Richard Blumenthal say the public should not be forced to subsidize data center bills. They are calling for stronger oversight and new rules for facilities that require massive power contracts.

Community frustration is also rising. Some residents do not want more warehouses full of servers that hum loudly and drive up neighborhood bills. In places with high density clusters, the resentment has begun to shape local elections.

A grid that is struggling to keep pace

The PJM grid operator which covers Virginia, Ohio and Illinois has faced a huge imbalance between supply and demand. Prices for capacity auctions, the mechanism used to make sure the grid can reliably meet demand, exploded this year. Bills jumped from two billion dollars to fourteen billion dollars in a single auction cycle and then climbed again to more than sixteen billion dollars.

Independent analysts say data centers account for over half of the projected demand costs in the region. That figure shows how rapidly AI related growth has transformed the energy market.

Other states offer a different picture. Texas has more than four hundred data centers yet saw only a modest increase in electricity prices. California has some of the highest electricity prices in the country, but its year over year increase barely nudged upward. Local conditions and grid structures play major roles in how data centers influence cost.

The future does not look cheaper

Most experts see little chance of electricity prices falling soon. The grid needs huge upgrades. Renewable energy projects wait years to connect. Transmission lines cost more to build than ever. Meanwhile AI companies announce new data centers almost every month.

The result is a new kind of techlash. AI has become the spark for political fights over who should pay for the infrastructure that fuels digital growth. Voters want relief. Politicians want answers. Utilities want more supply. And the tech industry wants more power to keep expanding.

No one expects the pressure to ease any time soon.