Marc Andreessen, Ben Horowitz and OpenAI president Greg Brockman are among the backers of an advertising campaign about to land on televisions in Ohio, Wisconsin and Kansas, telling voters there that AI data centers are good for them. Bloomberg reports that the group behind it, Build American AI, is spending millions of dollars; it will not say how many. The states were chosen because they are close ahead of the November vote, and the spend arrives weeks after national polling on data centers hit its worst reading yet for the industry.
That reading is the whole context for this campaign. An August survey by the Annenberg Public Policy Center at the University of Pennsylvania found a record 61% of American adults opposed to new data centers being built near their homes, up 12 percentage points from the same question in March. Twelve points in five months is not a soft trend that needs correcting at the margins. It is a majority forming in public, quickly, without anyone spending money to form it.
The money on the other side is organised. The campaign is reportedly financed by Leading the Future, a super PAC with a $100 million budget, and the named backers are people with direct exposure to whether the buildout continues: two venture capitalists whose portfolio depends on it and the president of the company that has done more than any other to make data center demand a national issue.
The message is where it gets strange. In a statement to Bloomberg, a Build American AI representative framed the data center buildout in existential terms for the country, and repeated the claim that opposition to these facilities amounts to a radical psychological operation, a position that happens to be extremely popular with American voters. The group said the stakes are too high to let the loudest and most radical voices decide America's future, and that turning away responsible investment would cost well-paid jobs, weaken local communities and stop the country from holding its lead.
Set those two things next to each other. Sixty-one percent of adults hold the view, and the campaign describes the view as a psyop run by radicals. This reads less like persuasion than like a group that has not yet accepted the poll is real. You do not buy airtime in three states to talk a psychological operation out of existence; you buy it because you think the people watching can be moved. The framing and the media plan are arguing with each other.
The choice of states says what the money is actually for. Ohio, Wisconsin and Kansas are named as places where the fight for votes matters, not as the places with the most data center construction. This is a campaign aimed at a ballot, not at a neighborhood, and the argument it makes fits that: it is built almost entirely around what the country forfeits by not building, with the local community appearing as a beneficiary of a decision taken somewhere else. That is a national argument delivered to people who are being asked a local question.
The group discloses its donors and withholds its budget, which is an odd way around for a campaign whose case is that the public has been misled about who is really behind the opposition. "Millions of dollars" is the only figure it offers, against a number it cannot dispute.
A few dramatic television spots are unlikely to repair an image this damaged before November. The more interesting outcome is what the industry concludes if the ads fail. A group sitting on $100 million does not stop at the point where the television budget runs out, and the spending that follows will not be aimed at changing minds.