Bring Back Nuclear

Three words on a retreat agenda. No question mark this time. No coy parenthetical. Just an imperative: Bring Back Nuclear.

According to WIRED's verified reporting, this is one of the scheduled sessions for Dialog's August 2026 retreat at the Powerscourt Hotel outside Dublin — the same gathering where 222 of the world's most powerful people will also discuss navigating World War III, battlefield technologies, and how to build a cult. The nuclear session stands out because, unlike most of the other agenda items, this one is already happening. Nuclear energy is coming back. The question Dialog appears to be discussing isn't whether to do it. It's how to control the terms.

And as it happens, the people in that room have very specific reasons to care about the answer.

---

How Nuclear Died

Nuclear energy didn't die because of physics. It died because of narrative.

On March 28, 1979, a partial meltdown at Three Mile Island Unit 2 in Pennsylvania became the worst nuclear accident in American history. No one died. Radiation releases were minimal. The Nuclear Regulatory Commission later concluded that the health effects on the surrounding population were negligible. None of that mattered. What mattered was the imagery: hazmat suits, evacuation orders, the word "meltdown" entering the American vocabulary as a synonym for catastrophe. Public support for nuclear energy, which had been climbing through the 1970s, reversed overnight.

Seven years later, on April 26, 1986, Reactor No. 4 at the Chernobyl Nuclear Power Plant in Soviet Ukraine exploded. Chernobyl was everything Three Mile Island wasn't — a genuine catastrophe with a confirmed death toll, a permanent exclusion zone, and radioactive contamination spread across Europe. The disaster became a symbol not just of nuclear danger but of institutional failure. The Soviet government initially tried to hide the accident. Western European citizens learned about the contamination before the affected populations in Ukraine and Belarus did.

Then came Fukushima. On March 11, 2011, an earthquake and tsunami disabled the cooling systems at the Fukushima Daiichi plant in Japan, triggering three reactor meltdowns. Again, the direct death toll from radiation was zero — the World Health Organization concluded that anticipated health effects from radiation exposure were low. But the psychological and political impact was enormous. Germany announced a complete nuclear phase-out within months. Japan shut down its entire reactor fleet for inspection. Global public support for nuclear energy dropped to its lowest point in decades.

Three accidents across four decades. Two of them killed no one from radiation. The third killed fewer people than a single year of coal-plant air pollution in any medium-sized country. And together, they effectively halted the global expansion of the most energy-dense, lowest-carbon baseload power source ever engineered.

Nuclear energy didn't die from its failures. It died from the gap between its actual safety record and the public's perception of its danger — a gap that was never successfully closed because no one with sufficient resources had sufficient incentive to close it.

Until now.

---

What Changed

What changed is artificial intelligence, and the staggering volume of electricity it requires.

Training a single large language model can consume as much electricity as several thousand American homes use in a year. Running inference — the process by which an AI model responds to queries — multiplies that demand across billions of daily interactions. Data centers, which already consumed roughly 460 terawatt-hours globally in 2024, are projected to reach 1,300 terawatt-hours by 2035, according to industry estimates. That growth rate is roughly equivalent to adding the entire electricity consumption of Japan to global data center demand in a single decade.

Solar and wind can't meet this alone. They're intermittent. A data center running AI workloads needs uninterrupted baseload power — the kind that doesn't disappear when the sun sets or the wind stops. Natural gas works but produces carbon emissions that tech companies have publicly pledged to eliminate. Hydroelectric is geographically constrained.

Nuclear is the only existing technology that provides carbon-free, 24/7, high-density baseload power at the scale AI demands. And the tech industry figured this out almost simultaneously.

In September 2024, Microsoft signed a 20-year, $16 billion power purchase agreement with Constellation Energy to restart — of all places — Three Mile Island's Unit 1 reactor, which had operated safely for decades before being shut down for economic reasons in 2019. The rechristened Crane Clean Energy Center is expected to produce 835 megawatts of clean power annually, all of it directed to Microsoft's AI operations. The U.S. Department of Energy backed the project with a $1 billion loan.

Google signed a deal with startup Kairos Power for 500 megawatts of power from small modular reactors. Amazon invested over $20 billion to convert the area around Pennsylvania's Susquehanna nuclear plant into an AI campus and backed nuclear startup X-energy with $500 million. Meta signed a 20-year power purchase agreement with Constellation for 1.1 gigawatts of nuclear energy from an Illinois plant, and issued a request for proposals seeking an additional 1 to 4 gigawatts of new nuclear generation. Oracle announced plans for a gigawatt-scale data center powered by three small modular reactors.

In aggregate, Big Tech signed commitments for over 10 gigawatts of new U.S. nuclear capacity in a single year. The industry went from indifferent to all-in faster than any energy transition in modern history.

Energy Secretary Chris Wright — visiting the Three Mile Island site in December 2025 — called the restart a poster child for the Trump administration's energy and AI agenda. In May 2025, the president signed an executive order titled "Reinvigorating the Nuclear Industrial Base," explicitly linking nuclear expansion to what it called a global race to dominate artificial intelligence.

Nuclear is back. Not because the public changed its mind about safety. Not because anyone resolved the waste storage question. Not because the disasters of 1979, 1986, and 2011 were reassessed. Nuclear is back because the most powerful technology companies on Earth need it, and they need it now.

---

Follow the Money Into the Room

Here is where the Dialog retreat agenda stops being abstract.

Peter Thiel, Dialog's co-founder, is not merely a bystander to the nuclear renaissance. His venture capital firm Mithril Capital was an early investor in Helion Energy, a nuclear fusion startup that has raised over $2.2 billion. The lead investor and chairman of the board for Helion's major funding rounds? Sam Altman, CEO of OpenAI — the company whose chief strategy officer appears on the Dialog registration list, according to WIRED. Altman is also chairman of Oklo, a publicly traded advanced nuclear fission company developing small modular reactors. He personally led Oklo's path to market through a SPAC merger.

Helion's investor roster reads like a Dialog pre-party: Altman, Thiel's Mithril, Facebook co-founder Dustin Moskovitz, and Reid Hoffman — Dialog's other co-founder.

Meanwhile, Palantir Technologies — co-founded by Thiel, with Dialog registrant Joe Lonsdale listed as co-founder — holds a $89.9 million contract with the National Nuclear Security Administration and recently partnered with The Nuclear Company, a reactor construction startup, in what Palantir called its first major foray into nuclear power. Palantir's software platform is now being positioned as a construction management tool for building new nuclear reactors, in addition to its existing contracts with the Pentagon, ICE, and the intelligence community.

Dialog registrant Scott Bessent serves as Treasury Secretary in an administration that issued four executive orders expanding nuclear energy in May 2025 alone, including invoking the Defense Production Act to strengthen the nuclear fuel supply chain. The Department of Justice, whose antitrust division would normally scrutinize industry-wide coordination on fuel supply, approved the consortium agreements.

The tech companies making the largest nuclear commitments — Microsoft, Google, Amazon, Meta — all have executives, board members, or close associates on the Dialog registration list. The government officials setting nuclear policy are in the same room. The venture capitalists funding nuclear startups are in the same room. The defense contractors building nuclear-adjacent software platforms are in the same room.

The session isn't called "Discuss Nuclear" or "Evaluate Nuclear." It's called "Bring Back Nuclear." The verb is active. The direction is settled. What's left to discuss is execution — and these are the people positioned to execute it.

---

The Bipartisan Miracle

One of the more remarkable features of the nuclear renaissance is that it has achieved something almost nothing else in American politics has managed: genuine bipartisan support.

According to Pew Research Center polling from 2025, roughly 60 percent of U.S. adults now favor building more nuclear power plants, up from 43 percent in 2020. Support has increased across party lines. The 2024 ADVANCE Act, which streamlined nuclear licensing and provided financial incentives for advanced reactor development, passed both chambers of Congress with overwhelming bipartisan majorities. According to the American Nuclear Society, the 2024 Democratic platform endorsed nuclear energy for the first time since the Nixon administration — meaning that for the first time in over 50 years, both parties officially support nuclear power.

This is the kind of consensus that typically indicates either a genuinely good idea or a well-coordinated lobbying effort. It can, of course, be both.

The nuclear industry spent heavily on lobbying in the years preceding this alignment. Tech companies poured resources into both parties. The policy outcome — streamlined permitting, federal loans, executive orders, Defense Production Act invocations — aligned perfectly with the commercial interests of companies whose executives are registered for the same off-the-record retreat in Dublin.

Correlation is not causation. But when a bipartisan legislative outcome coincides precisely with the financial interests of the people who attend the same private, off-the-record gathering, it is at minimum worth noting who was in the room before the consensus formed.

---

The Right Answer for the Wrong Reasons

Here is the uncomfortable part: the nuclear advocates might be right.

The climate math is stark. Coal kills an estimated 800,000 people per year through air pollution, according to research published in Environmental Research. Natural gas is cleaner but still emits significant CO2. The International Energy Agency projects that global electricity demand will grow by roughly 75 percent by 2050. Meeting that demand with renewables alone would require a buildout of solar, wind, and battery storage at a pace never achieved in human history — and even then, the intermittency problem remains unsolved at scale.

Nuclear energy produces zero direct carbon emissions during operation. Modern reactor designs are substantially safer than the plants that failed at Three Mile Island, Chernobyl, and Fukushima. Per unit of energy produced, nuclear has the lowest death rate of any power source — lower than wind, lower than solar, dramatically lower than fossil fuels. The fuel is energy-dense: a single uranium pellet the size of a fingertip contains as much energy as 17,000 cubic feet of natural gas.

Small modular reactors, if they achieve commercial viability, could bring nuclear capacity to locations and scales that were never practical with conventional gigawatt-scale plants. Factory construction could reduce costs and timelines. Standardized designs could eliminate the bespoke engineering that has historically made nuclear projects late and over budget.

This is not fringe thinking. It is the emerging consensus of climate scientists, energy economists, and policy analysts across the political spectrum. Nuclear energy is, by most credible analyses, an essential component of any realistic decarbonization pathway.

So the question isn't whether nuclear should come back. The emerging evidence suggests it should. The question is: who benefits from the way it comes back?

If the nuclear renaissance is structured so that tech companies sign 20-year power purchase agreements that lock in clean energy for their AI operations while surrounding communities see no rate relief and no grid benefit, then "Bring Back Nuclear" means something very specific. It means: bring back nuclear for us.

If Three Mile Island's 835 megawatts go to Microsoft data centers under a private contract while Pennsylvania consumers in the PJM electricity market face rising prices from AI-driven demand, then the restart isn't an energy policy. It's an energy acquisition.

If the permitting reforms and federal loans flow primarily to projects with guaranteed offtake from trillion-dollar tech companies, while independent nuclear developers and public utilities struggle to access the same capital, then "Bring Back Nuclear" is a market structure choice, not just an energy choice.

---

The Architecture of Inevitability

The Dialog retreat agenda doesn't phrase it as a question. There's no hedge, no parenthetical, no "does?" — just the imperative. Bring it back. And the composition of the room — the venture capitalists who funded the startups, the executives whose companies signed the power contracts, the government officials who signed the executive orders, the defense contractors positioned to build the infrastructure — suggests that the discussion isn't about whether to proceed. It's about optimizing the terms.

This is what coordinated action looks like when it doesn't need to be conspiratorial. Everyone in the room has aligned incentives. The tech companies need power. The investors hold nuclear assets. The policymakers want the economic activity. The defense establishment sees strategic value. Nobody needs to collude. They just need to be in the same room, talking off the record, with no FOIA trail — which is, according to WIRED's reporting, exactly what happens at every Dialog retreat, and exactly how every registered government official arranged their attendance.

Nuclear energy is probably the right answer. The troubling part isn't the answer. It's the process — a process in which the people who profit from the policy, the people who make the policy, and the people who enforce the policy all meet behind closed doors, and the rest of us learn what they decided after the contracts are signed.

Bring back nuclear. But notice who's doing the bringing, and ask what they're bringing it back for.

---

Sources

---

This is original analysis published by Build a Cult. It is licensed under CC BY 4.0. The session title "Bring Back Nuclear" is drawn from Dialog's leaked 2026 retreat agenda, as reported by WIRED.