In Ottawa, Kansas, residents petitioned for a November 3 vote on whether to prohibit large data centers designed to draw 25 megawatts or more. They are worried about their electricity, their water, and who gets to decide how those resources are used.
A Virginia super PAC called Defend U.S. responded by sending mailers urging residents to vote against the ban. The locals have been going door to door. As one organizer told KSHB, they cannot afford the billboards and mailers.
Here is the contest in miniature: people trying to govern their own town, and outside money trying to govern the people.
Elsewhere in Kansas, a separate pro-data-center super PAC, Building the Future, is spending more than $1.5 million to help elect Republican gubernatorial candidate Ty Masterson. It grew out of Leading the Future, a national pro-AI political group. A related nonprofit, Build American AI, says it has roughly $50 million available for advocacy in Kansas, Ohio, and Wisconsin. NBC News and Axios reported the campaigns.
That $50 million is available cash, not money already spent. The Ottawa mailers come from a different PAC. None of those distinctions changes the larger picture: AI infrastructure has become politically important enough to warrant an organized campaign for the politicians and policies its backers prefer.
A building full of computers now comes with an election strategy.
The cloud has a landlord
“Cloud” was a brilliant choice of name for something that needs a power plant.
For two decades, the technology industry taught us to imagine computing as weightless. A few clicks and the machinery disappeared. Now even intelligence appears to materialize inside a chat window, without land, labor, water, electricity, or a bill for any of it.
The largest AI campuses are putting an end to that illusion. A June 2026 report from Lawrence Berkeley National Laboratory projects data centers could consume 9.5% to 15.3% of all U.S. electricity by 2030, depending on the scenario. The central estimate is 11.8%. Forecasts are uncertain, but utilities have to make construction decisions years before the bill arrives.
Fanny Lopez and Cécile Diguet's Digital Conquest examines the territorial reality of computing: private companies absorbing land, energy, and space while governments and communities manage the consequences. Their account makes the word cloud seem increasingly absurd. Every digital service sits somewhere, and somebody else was already living there.
In Hyperscale, Paris Marx examines the giant computing empires being assembled around AI. His reporting on water conflicts brings the argument down to earth, often in places where water is already contested.
A data center can be economically useful. A town can reasonably want the tax revenue and the jobs. But the decision belongs to the town as much as it belongs to the developer. That elementary proposition is apparently controversial enough to require political advertising.
Privatize the upside. Negotiate the public bill.
The economic pitch is simple. Offer favorable terms now; prosperity will follow.
Sometimes it does. Construction employs skilled workers. Data centers pay property taxes and support parts of the wider economy. The question is whether the benefits justify the concessions and long-term obligations being signed away.
A Brookings study published in 2026 found that a typical county receiving its first major data center gained roughly 100 to 200 jobs over time, depending on the facility. Wages barely moved. The researchers also found that some employment estimates overstated the benefits by comparing data-center counties with places that weren't growing as quickly before development began.
Meanwhile, Virginia's sales-tax exemption for data centers cost an estimated $1.6 billion in fiscal 2025. That's a great deal of public generosity for an industry whose leading firms are among the wealthiest enterprises on the planet.
Then there is the electricity system. Who pays for generation, substations, and transmission upgrades? Who absorbs the risk if demand never matches the forecasts? What happens when household bills rise while the largest new customer enjoys a specially negotiated rate?
These aren't objections dreamed up by people who dislike technology. They're the questions any serious investor would ask before committing capital. Somehow, when the public is the investor, asking them becomes an obstacle to progress.
The danger is not that every data center is a bad deal. It is that the people negotiating on behalf of the public may have less information, less bargaining power, and far less money to shape the political outcome than the companies seeking approval.
The billionaires have bought a megaphone
The political money deserves names.
The San Francisco Chronicle reported that OpenAI president Greg Brockman and his wife, Anna, donated $25 million to Leading the Future and $25 million to MAGA Inc., a Trump-aligned super PAC. Andreessen Horowitz has also backed Leading the Future, according to TechCrunch.
OpenAI says it has made no corporate PAC donations and that Brockman's political contributions were personal. That is the formal distinction. It does not erase the political influence of a $25 million contribution by the president of a company standing to benefit from AI-friendly policy.
None of this requires a secret conspiracy. The donors, the campaigns, and the financial interests are largely in plain sight. That's precisely what makes it so brazen.
The same people who finance the infrastructure want influence over the rules governing it. They can afford nationwide advertising campaigns, armies of lawyers, and access to elected officials. A town in Kansas can afford to knock on doors.
We call that a debate about the future. The participants have rather different budgets.
The ideology is doing its job
The money is accompanied by an argument: technological acceleration is not merely desirable but morally obligatory.
Marc Andreessen made that argument in his Techno-Optimist Manifesto. He asserted that slowing AI would cost lives and described deaths that could have been prevented by AI as a “form of murder.” Under that reasoning, opposition to his preferred pace of development begins to look less like democratic disagreement and more like a moral crime.
His manifesto also places sustainability, social responsibility, trust and safety, and technology ethics among ideas he considers enemies of progress. It's an extraordinary way to talk about a technology whose developers are asking entire communities to trust them.
In August 2026, Andreessen Horowitz announced a $1.1 billion Machine Age Fund to invest in the hardware and infrastructure behind AI. The firm called rapid physical expansion a “social and national imperative.”
The philosopher might call that a worldview. The investor might call it a thesis. The people paying for a new substation may have another name for it.
Paris Marx's essay on longtermism, excerpted from Hyperscale, explains the intellectual convenience of appealing to humanity's distant future. If trillions of hypothetical future lives depend on AI reaching some imagined destination, the material interests of people alive now can be made to look petty by comparison.
The future is a particularly cooperative witness. It never attends a zoning meeting to ask why the water is gone.
Gil Durán's The Nerd Reich investigates another strand of the same elite milieu: the attraction to forms of power that dispense with democratic accountability. Peter Thiel had already written in 2009 that he no longer believed freedom and democracy were compatible. These are not the ideas of everyone building AI. They are the published ideas of people with substantial influence over the technology economy.
The old Silicon Valley mythology celebrated rebels who gave ordinary people access to powerful tools. Some of its richest heirs now seem to regard ordinary people using political power as a design flaw.
The problem is not ambition. I admire ambition. The problem is the entitlement that turns an investment strategy into a claim of authority over other people's lives.
More intelligence. Less agency.
I want AI to succeed. I want powerful tools, new science, cheaper expertise, better medicine, and technology that gives individuals and small organizations the capabilities of much larger ones.
Those possibilities are worth building for. They are also worth defending against the people who would turn them into another system of dependence.
The danger is easy to miss because the technology itself can be remarkable. A business gains an AI assistant, then discovers its knowledge, operations, and costs are tied to a provider it cannot easily replace. A community gets a shiny new campus, then learns how much of its future electrical capacity has already been committed. A country celebrates domestic AI investment while a small number of private owners acquire more influence over its energy plans and lawmakers.
The computers are becoming more capable. Who is becoming more powerful?
Ownership of computing infrastructure carries the power to set prices, dictate access, decide which competitors can afford to enter, and shape the terms on which others participate. Political influence can reinforce that ownership. Control over infrastructure then provides the wealth to buy still more influence. It is a feedback loop with very human consequences.
The alternative is not a ban on computing or a server rack in every home. It's a technological economy with more choices: open models, portable knowledge, independent providers, public oversight, fairly priced infrastructure, and communities with a real voice in the projects built around them. Lopez and Diguet examine different possible digital futures, including decentralized public infrastructure. The current arrangement is neither inevitable nor sacred.
Data centers will help determine who possesses the means of producing and using advanced intelligence. That is an extraordinary amount of power to allocate through a mixture of private contracts, public subsidies, and political spending.
A handful of billionaires can finance the machines. They do not thereby acquire the right to design the society around them.
We need more intelligence, not a new aristocracy that owns it.