Nate B Jones opens with the claim that the AI data center buildout "had a terrible night" on election night, November 3rd, 2026, when communities voted to block or restrict data centers. His point is that this is not a video about data centers alone. It is about the second and third order effects when the compute infrastructure behind the new economy is constrained. He argues AI use is not slowing, so the interesting question is what the constraint does to companies, communities, and the odds of an infrastructure bubble.
What it is. Jones cites a May Edison Research survey finding 65% of American adults had used an AI chatbot in the previous week, up from 52% in February. He also cites an October ABC News, Washington Post, and Ipsos poll finding 72% of Americans opposed a new data center in their community. In Edgewater, Florida, he says unofficial results showed more than three quarters of voters backing a prohibition on new data centers, and he calls those typical margins.
Why it matters. He says Americans seem to expect AI to come without data centers, "out of the air," and that it does not. He notes Google said its model API usage was up 40% quarter over quarter and more than four times a year earlier, and that Google had begun using outside compute capacity because of supply constraints on Google Cloud.
How to apply. Do not read the vote as a sign that demand is falling. Treat use and infrastructure opposition as two separate signals.
What it is. Jones describes neoclouds, specialist cloud companies that rent out powerful AI chips. If they planned to expand quickly on low-regulation Texas sites, longer approvals could make the business unviable. He lists the pressure: borrowed money, equipment already committed, a customer expecting capacity by a set date, and equipment that keeps depreciating.
Why it matters. He argues the company that fails does not take the project with it, because someone with cash will buy it. Existing data centers gain from fewer competing facilities, which can support higher utilization, better pricing, and stronger negotiating power. In his words, the backlash "may end up, ironically, strengthening big tech," and smaller AI companies may end up renting capacity from the big players they compete with.
How to apply. If you run or buy from a smaller AI company, ask who controls its compute. Access to computing becomes a bigger lever in whether a product is viable.
What it is. Jones says Derek Thompson made the argument after the election that slowing supply gives demand more time to catch up. He describes developers who all read the same growth forecast and each decide to build, so that every project looks reasonable but together they build far more capacity than the market needs.
Why it matters. He stresses we have not yet seen that glut, but it was a risk he was worried about for 2027. Moderating the U.S. buildout could reduce the chance of a supply glut over the following year or two, even while it creates serious problems for individual developers.
How to apply. Keep two ideas in mind at once: the effect on any one developer can be severe while the effect on the whole economy may be to avoid a bubble. Jones says he is not predicting what any stock or neocloud is worth.
What it is. Jones frames the story as demand that has been blocked. A company that needs more compute next year will not wait for the next legislative session. It looks elsewhere, perhaps a different cloud provider or a location that looked more expensive but has deliverable power, water, and community support.
Why it matters. He says a developer who can open on time offers something different right now, and that this shifts bargaining power toward communities, who can negotiate over infrastructure costs, union jobs, and local investment. He adds that the usual map of friendly and unfriendly states is less useful, and that you have to look at specific sites, power agreements, and local support.
How to apply. When evaluating a compute plan, look at the site level instead of the state level.
What it is. Jones says some displaced projects will go outside the U.S. He mentions Europe's Gigafactory initiative and a U.K. agreement on what it expects from developers, but says he hears more interest outside the EU because building in the EU is not seen as much faster. He names Canada and Mexico as interesting, and says India, Southeast Asia, and parts of South America may gain.
Why it matters. Jobs and spending follow the investment, while the AI service gets "effectively rented back" to American customers. His line: "If you keep a data center out of the community, it doesn't keep the technology out. It just means you're renting the technology and you don't own it."
How to apply. Assume an uncertain U.S. political environment for at least the next two years and plan site choices state by state, and beyond the U.S. where needed.
What it is. Jones calls this perhaps the most speculative bet of 2026. He notes xAI has said publicly it is pursuing orbital compute, that Elon and Jensen have talked about it, and that Google has a Project Suncatcher prototype in orbit.
Why it matters. He separates a prototype from solving radiation, thermal, depreciation, and maintenance problems, but says orbit has a much looser regulatory environment. He says the election raises the odds of attacking those problems with more capital. If he were betting, his estimate would move from 10 years to maybe 6 to 8, and he says that is not a done deal.
How to apply. Treat this as his personal estimate, not a forecast to plan around.
What it is. Jones suggests watching what companies do with the next project they planned. Are they waiting, moving, paying more for a deliverable location, buying capacity from someone else, or investing somewhere without the same constraints?
Why it matters. He says those decisions will tell you more about a company's commitment to AI than a press release, and he likes reading where people allocate capital on the balance sheet. For your own work, he cautions against using construction headlines as a forecast for adoption. Your employer will keep buying AI services, the price may rise a bit, and available suppliers may change.
How to apply. Follow capital allocation, and keep adoption plans tied to demand rather than to data center headlines.
What it is. Jones argues the rebuke of data centers came from concern about AI's place in the American Republic. He says many Americans feel AI is set against the American dream and is positioned as disempowering, and that "Silicon Valley has itself to blame on this." He also says talk of making humans extinct is not winning.
Why it matters. He shares a note from a tech friend at an AI company who worries about his job. He says the industry must be clear about benefits, why AI adds value to your day and does not just automate away your job. He adds that Europe is not very optimistic on AI either, and that fragmenting around local fights over infrastructure sites would be a tragedy.
How to apply. Talk about empowerment, tooling, training, and preparing the workforce. He says this is the question we can shape with storytelling.
Jones's answer to the title is that the boom is not ending, but it is being reshaped. Demand keeps growing while supply gets harder to site, which favors big, funded operators, rewards reliable sites, pushes some building abroad, and may reduce the risk of overbuilding. He closes by inviting viewers to share where they think AI is going after November 3rd, 2026.