Kevin O'Leary's Utah data center fight is really about who pays for the AI build-out

Kevin O'Leary wants to build a very large data center in Utah, and the people who actually live where it would go would prefer he didn't. That, in its plain form, is the story. Everything else — the China accusation, the political theatre, the celebrity-investor shtick — is garnish on a question that the AI build-out is now forcing on every county in the American interior: who absorbs the costs of the next industrial cycle, and who decides.

O'Leary, the Canadian businessman and Shark Tank fixture turned AI infrastructure pitchman, has framed the local resistance as something other than a zoning fight. According to reporting circulating on 14 June 2026, he has been telling interviewers that China is stoking the opposition — that Beijing-linked actors are funding or amplifying the NIMBY campaign against his Utah project, because the United States cannot be allowed to win the AI capacity race. The claim is striking, and the evidence for it, on the public record, is thin. Local opposition to hyperscale data centers in the Mountain West is well-documented, well-organised, and has perfectly mundane causes: water draw in a drought-stricken basin, noise and visual blight on open range land, grid strain, property taxes that do not begin to cover the public cost of new transmission lines, and a deep suspicion that the electricity and the cooling water will be sold to the highest bidder while the locals get the diesel backup generators.

The locals have a case, and it isn't made in Beijing

The most credible driver of resistance to large AI campuses in Utah, Arizona, New Mexico, and west Texas is energy. Hyperscale data centers are the most electricity-hungry buildings ever built, and they cluster where land is cheap, the climate is cold, and the grid has surplus — which, in 2026, increasingly means the interior West. But surplus in the abstract is not surplus at the meter. When a single facility draws more power than a mid-sized city, the rate base shifts. Utilities pass costs on. Industrial customers with long-term contracts can lock in low tariffs; residential ratepayers, farmers, and small businesses absorb the marginal cost of the new generation and the new transmission. The pattern is familiar from the shale boom of the 2010s, when mineral rights holders got rich and the counties hosting the wells paid for the road repair.

O'Leary's accusation is the latest iteration of an old reflex in American industrial politics: when local opposition to a project threatens the timeline, blame a foreign adversary. The rationale is that security framing unlocks federal permits, fast-tracks environmental review, and discredits opposition as tainted. It is a move that has been made on LNG terminals, on semiconductor fabs, on rare-earth processing, and on solar farms alike. It sometimes works. It does not require the foreign-adversary claim to be true, only to be plausible enough to occupy a news cycle.

The China angle is real, just not the way O'Leary means

That said, the China frame is not empty. Beijing has spent the last five years building its own AI compute base, subsidising domestic chip design and fabrication, and racing the United States for the inputs — advanced lithography, high-bandwidth memory, substation-scale transformers — that hyperscale data centers depend on. The contention between Washington and Beijing over semiconductor export controls, advanced-node fabrication, and the routing of Nvidia accelerators is the most consequential industrial-policy story of the decade. State-on-state competition in compute capacity is genuine.

What is not genuine, on the available evidence, is the claim that this competition explains a specific NIMBY fight in a specific Utah county. Local opposition to data centers predates the latest export controls. It predates the LLM build-out. It is structurally similar to the opposition that greeted wind farms in upstate New York, transmission lines in Maine, and crypto mining in upstate New York and the Texas hill country. The pattern is the same: a capital-intensive project arrives, promises jobs, and asks a rural community to underwrite the externalities. Sometimes the community says yes. Increasingly, it is saying no, with lawyers.

What O'Leary is actually selling

The most useful way to read O'Leary's intervention is not as a property dispute but as a marketing move. O'Leary is, in 2026, more an asset-management brand than a television personality. He pitches AI and crypto exposure to retail investors, runs special-purpose acquisition vehicles, and takes board seats at companies that need a recognisable face. A data center is a long-duration, capital-intensive asset class that fits his pitch perfectly — steady contracted revenue, infrastructure-grade yields, a story about the future. The China angle in the press tour serves the same function: it elevates a county-level permitting fight into a national-security narrative, which lifts the asset class, which lifts the pitch.

The bet is that voters and regulators who would happily resist a speculative real-estate play will fold in the face of a China frame. It is a reasonable bet. It is also a sign of how thin the public tolerance for the AI build-out's externalities has become, that a celebrity pitchman feels compelled to reach for a foreign adversary to sell it.

Stakes

If O'Leary's framing succeeds, expect the next wave of hyperscale projects to be justified in security terms rather than commercial ones, and expect the costs — water, grid capacity, transmission — to be socialised under the same logic that built the interstate system and the rural electrification programme. If it fails, expect a slower, more contested build-out, more local vetoes, more state-level negotiation over rate-base allocation, and a more honest conversation about who pays for the compute that the rest of the economy is rapidly coming to depend on.

The locals in Utah are not pawns in a US-China duel. They are the ratepayers who will pay for the transformer, the water rights, and the substation. Treating that as a foreign-influence problem is the most American resolution available, and probably the worst one.

This publication finds that the China-blame frame around O'Leary's Utah project is, on the public record, unsupported by anything more than a talking point — and that the underlying fight is a familiar one about who underwrites industrial booms in the American interior. We will update if reporting surfaces evidence to the contrary.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/cointelegraph
• https://t.me/cointelegraph
• https://t.me/cointelegraph