Gov. Kathy Hochul’s first-in-the-nation moratorium on large data centers carries one big lesson: Artificial intelligence has a politics problem that no amount of lobbying will fix.
AI is one of the most valuable technologies ever to emerge, yet it’s increasingly losing the political battle for public support.
Without that support, America will squander its huge advantages and fail to benefit from what it’s built.
That’s why AI’s advocates need to consciously cultivate a pro-AI majority — and fast.
The polls tell the story: New Yorkers backed a data-center moratorium by better than two to one in last month’s Siena survey.
Nationally, seven in ten Americans now oppose a data center in their area, which is up roughly 30 percentage points in under a year.
Hochul, in other words, isn’t defying public opinion; she’s following it.
The risk of course is that more restrictions could be coming, including punitive taxes, permitting refusals and local vetoes — all aimed at the most consequential technology of the era.
But good news: The raw material for a pro-AI majority exists.
Assembling it, though, requires thinking differently about political coalitions — because AI doesn’t divide the country along party lines.
Recent polling from Echelon Insights maps the coalition.
Two groups form its natural base: Young, tech-enthusiast voters and optimistic center-right Republicans together represent about 30% of the electorate.
Add college-educated suburbanites who like technology but remain lukewarm on AI specifically, and it reaches roughly 42%.
The swing bloc is the largest group of all — younger, mostly female voters who are skeptical of AI but not committed against it.
Win a share of them, and a majority is in reach.
The key point here is that a pro-AI majority coalition can be built without flipping outright opponents.
It’s about reaching technological optimists who haven’t yet heard a reason to extend that optimism to AI itself.
Giving them one requires a better story than Silicon Valley boosters have provided, and a positive narrative that counters opponents’ doomsaying.
The honest story happens to be the best one: AI spreads capability downward.
The research is remarkably consistent.
For example, customer-service agents using an AI assistant resolved 14% more issues per hour — but the gain was 34% for the least experienced workers, and nearly zero for the stars.
Studies of writers, coders and consultants find the same pattern: The biggest boost goes to those starting furthest behind.
In practice, that’s a Queens contractor getting a lawyer’s help with a payment dispute for the price of a monthly subscription.
It’s a bodega owner with a bookkeeper that never sleeps, or an immigrant with an idea building a software program that once required a funded engineering team to produce.
AI is the most powerful solvent yet invented for erasing the barriers to starting things.
Call the resulting agenda “diffusionism,” with the goal of getting AI’s capabilities into as many hands as possible.
The point is not just economic growth, but the broad distribution of capability itself.
Most of the current AI debates begin by asking who will capture the returns from AI, and how those returns should be redistributed.
But a society can’t redistribute prosperity it never generates.
A diffusionist message speaks to every group the coalition needs — markets and initiative for the right, access and broken gatekeepers for the center-left, and a renewed sense of personal agency for the young.
Diffusion runs on physical things: chips, electricity and, indeed, data centers.
Which brings the argument back to Hochul, because the grievances behind her moratorium are politically potent.
Yet her ban answers none of those grievances.
It builds no power plants, lowers nobody’s utility bill and creates not a single job.
Worse, it sends the industry’s billions to Virginia, Ohio or Texas.
The correct response is a deal: require data centers to pay for the power they consume or to generate their own, deliver visible benefits to host communities, and fast-track the new energy generation the grid needs anyway.
The deeper opportunity, though, is political, and it’s sitting unclaimed.
More and more politicians are lining up to run against AI.
The first leader to run for something instead — for the contractor, the bodega owner, and the founder; for cheaper power and broader access to the most capability-enhancing tools ever built — will find the largest bloc of persuadable voters in America waiting.
Hochul is betting that the backlash is where the votes are, through November and beyond.
But her pause expires within a year.
The politics of what comes after will belong to whoever has built the majority by then.
Jesse Arm is the vice president of external affairs at the Manhattan Institute, where Sean Speer is a technical writer. Adapted from City Journal.
