It is, on the face of it, an odd pairing: a man whose business history includes a topless cabaret in Youngstown, a 30-day jail stretch for selling beer without a licence, and a federal probe of a drug-treatment centre he ran in the California desert — paired with a 950,000-square-foot proposal to build what would be one of the largest data centres in California, on dusty lots in Imperial next to a Homeland Security complex, to be cooled by Colorado River water that does not belong to him. The pairing is, on closer inspection, the story.
Sebastian Rucci, 65, has preliminary approvals for the 330-megawatt facility. He says he has spent $5 million of his own money on it. He says he has interest from Google. Google says it has no involvement in the project and would not, in its position, want to be. The Imperial Irrigation District has refused to fallow farmland and re-allocate the water. The county has imposed a one-year moratorium. Rucci has sued the county, the irrigation district, a local activist, and a public-television station that covered the project. He has called a state senator who introduced bills to regulate data centres “a buffoon.” A 22-year-old from El Centro has been arrested for an online threat to kill him. A judge has dismissed two of his defamation suits. A high-school band director named Christopher Scurries lives behind the proposed site and has written a folk song about it.
It is a lot of story for a project that exists mostly on paper. The interesting question is not whether Imperial County should host a data centre — though that question is real, and the residents asking it are not wrong to ask it. The interesting question is why this particular project, at this particular site, with this particular principal, is the one that has managed to acquire preliminary approvals, opposition, lawsuits, death threats, and a folk song simultaneously, in a county that has spent the last year discovering what AI infrastructure actually costs the people who live next to it.
Let us walk through what the documentary record actually shows.
The plan, as reported, is for a 330-megawatt data centre — 950,000 square feet of compute, on the edge of the Sonoran Desert, in a region whose average July temperature runs north of 105°F. The cooling load of a facility of that size is not a footnote; it is the project. Rucci’s proposed path to the water was to purchase farmland, take it out of production, and use the Imperial Irrigation District’s allotment of Colorado River water to cool the servers. The IID turned him down in May. This is the part of the proposal that should have been the lede from the start, because the water math is the only part of a data centre proposal that cannot be papered over by a PowerPoint deck.
Imperial Irrigation District is the largest single user of Colorado River water — it holds rights, by its own accounting, to roughly 3.1 million acre-feet per year, more than the entire allocation of either Arizona or Nevada. That allocation has been the basis of the Imperial Valley’s agricultural economy since 1901, when the New York–financed irrigation works were completed and the desert became one of the most productive winter-vegetable growing regions in North America. The arithmetic of the project is therefore not “data centre versus nothing.” It is “data centre versus the lettuce and sugar beets and alfalfa currently grown with that water.” The IID’s refusal to allow fallowing and re-allocation is a refusal to convert an existing, documented, job-producing use of water into a use that produces neither food nor, in this case, a tax-paying tenant of any seriousness.
The project also promises, per Rucci, an estimated $28 million in annual tax revenue and 1,600 construction jobs. These numbers are presented in the manner such numbers are always presented in data-centre pitches — as a fait accompli, with no operating-cost counter, no electricity-rate counter, no estimate of who pays for the transmission upgrades, no estimate of the public-health costs of running a gas turbine peaker plant in a county that already fails the federal PM2.5 standard on bad-air days, which Imperial does. The 1,600 construction jobs are temporary. The few dozen permanent jobs such a facility would actually employ once built — to take one comparable data point, Vantage’s 1.1-million-square-foot Reno campus is on record at 73 permanent staff — are not mentioned in the pitch.
This is what data-centre politics looks like at the local level in 2026, and the pattern is consistent enough that the residents of Imperial are not, contra the developer’s framing, an irrational NIMBY flash mob. They are, in many cases, people who have been through this before. Christopher Scurries — the high-school band director who lives behind the proposed site and who googled the developer — is the kind of citizen-plaintiff who has become a structural feature of data-centre fights in the rural West: under-resourced, professionally untrained, technically outgunned by the developer’s consultants, and right about the underlying questions. He formed Not In My Backyard Imperial, organised the opposition, and wrote a folk song. The song is, by report, not subtle. The organisers behind recall campaigns in other Imperial Valley jurisdictions are not the same group, but the political infrastructure of resistance is being built along the same lines.
The Imperial County Board of Supervisors is, in its own telling, embarrassed. The same board that enthusiastically welcomed the project last year voted in June for a 45-day freeze and in July extended the freeze to a year. Rucci is suing over the moratorium. The Sierra Club has separately sued to force a full environmental review under the California Environmental Quality Act. The board’s political problem is straightforward: if it approves the project, it owns the water, the air, the noise, and the rate-base consequences when residents’ electricity bills go up to subsidise the transmission build-out. If it does not approve, it gets sued by a developer with a history of suing anyone who opposes him, and who has demonstrated both the resources and the appetite to do so.
Rucci’s litigation record is the part of the story that requires the most care, because the temptation is to treat it as a character study when the structurally interesting question is what it tells you about the cost of resisting a data-centre proposal. He has sued the county, the irrigation district, a local activist, and a public-television station. Two defamation claims have already been dismissed — one by a judge who wrote that “the defamation claim is not legally or factually supported.” These are not isolated moves. They are the standard playbook of a developer who has read the SLAPP-statute case law in California and concluded, correctly, that under the anti-SLAPP fee-shifting regime, the cost of defending even a dismissed suit is an order of magnitude greater than the cost of filing one, and that the deterrent effect on a county supervisor, a public-television editor, or a private citizen with a day job is large. The fact that Rucci has lost the suits he has filed does not make him less dangerous to the people he has filed them against. The fact that he is filing them is the message.
This is where the structure of the larger fight becomes visible. The local opposition in Imperial is not just opposing a data centre. They are opposing a developer who has demonstrated, in court filings already on the public docket, the willingness to use the legal system as a tool of attrition against the people who oppose him. The activist whom Rucci sued is exactly the kind of person the anti-SLAPP statutes in California were written to protect. The fact that the suit was dismissed is the system working. The fact that the suit was filed in the first place is the system being used.
The same dynamic is now visible in similar fights from Social Circle, Georgia to Utah to a growing list of jurisdictions where the residents’ first response to a hyperscale data-centre announcement has been to lawyer up. The developers, in turn, have responded with the inverse of the residents’ playbook: file suits, run political candidates, frame opposition as ideological, and wait. Rucci did all four. He ran a candidate for the IID board. He lost, but the cost to the incumbent of running a serious race in a small district is real, and the next developer who shows up in Imperial will find that the calculus of a board race is now part of the price of admission.
The political backdrop is the part the developers would prefer not to discuss. State Senator Steve Padilla of Chula Vista has introduced bills to regulate data-centre construction and to push the cost of grid upgrades onto the developers rather than onto ratepayers. This is, structurally, the same fight playing out in a dozen state legislatures — Virginia, Texas, Indiana, Ohio, Georgia — where the question is whether the ratepayers of the surrounding counties will pay the transmission and generation costs of facilities whose primary customers are large AI labs and hyperscalers operating in other states. Padilla’s bills are an attempt to answer that question in favour of the ratepayers. Rucci’s response to Padilla was to call him a buffoon. The response is, itself, a tell: when a developer calls a state senator a buffoon for introducing a bill that asks developers to pay for the grid upgrades their facilities require, the developer is telling you, in plain English, that the developer’s business model depends on someone else paying for the infrastructure.
The pattern, viewed from the outside, is consistent. AI buildout is a real-economy infrastructure programme with real-economy costs. The capital is concentrated in a small number of firms — Microsoft, Google, Meta, Amazon, with OpenAI and Anthropic as significant second-tier players — and the physical buildout is being located in rural and small-metro counties where the local government lacks the technical capacity to negotiate with the developer’s consultants and the political incentive to refuse the offered tax revenue. The cost of the infrastructure — water, electricity, transmission, air quality, road wear, noise — is borne by the residents of those counties. The benefit — the compute, the training, the inference, the eventual revenue — accrues to the firms whose customers are concentrated in other states and other countries. The local fight, whether in Imperial or in Loudoun County or in central Texas, is over who pays for what.
In this particular fight, the developer has a strip-club conviction record, an unsuitability for the work that is apparent to anyone who has toured an operating hyperscale facility, a litigation pattern that would be disqualifying in most regulated industries, and a 330-megawatt proposal on a site whose water, power, and air the local jurisdiction has every reason to believe it cannot afford to give him. The Imperial County Board of Supervisors has, with the moratorium, taken the only action available to a local government in this position. The IID has refused to fallow farmland. The Sierra Club is in court. A judge has dismissed two of the developer’s suits. A high-school band director has written a song.
These are not the actions of a community that has been captured by an irrational anti-data-centre ideology. They are the actions of a community that has read the developer’s filing, done the math on the water and the power, looked at the litigation record, and concluded, correctly, that the project on the table is not the project that was pitched. Whether the next project, with a different principal, on different terms, with a different relationship to the IID and the ratepayers, is the project Imperial is willing to host, is a question the residents have earned the right to answer on the record, and the one-year moratorium is the procedural mechanism by which they will answer it. Deadlines, as is sometimes noted, are the only part of regulatory processes that the regulated actually respect.
The work is to be done. The water is the work.