For decades, the message was clear: conserve. We were told to switch off the lights, buy energy-efficient appliances, and embrace a future powered by the gentle hum of wind turbines and the silent glow of solar panels. Nuclear energy, with its specter of mushroom clouds and radioactive waste, was the villain in this story—a relic of a bygone era, too dangerous to even consider. Proponents were dismissed as reckless, their calls for a powerful, carbon-free energy source drowned out by a chorus of environmental caution.
Then, almost overnight, the script flipped. The quiet hum of conservation has been replaced by the roar of a new industrial revolution, one demanding an almost insatiable amount of power. The villain has been recast as the hero, and the world is suddenly in a frantic, pedal-to-the-metal race to build a new nuclear age. But as the dust settles from this seismic shift, a chilling question emerges: a race to what?
The AI Energy Appetite
The driving force behind this nuclear renaissance is not a newfound public acceptance or a sudden breakthrough in waste disposal. It is the voracious, and frankly terrifying, energy appetite of artificial intelligence. The race to build Artificial General Intelligence (AGI)—and its even more powerful successor, Artificial Superintelligence (ASI)—has ignited a global arms race for computational power, and with it, a desperate scramble for the electricity to fuel it.
To understand the scale of this demand, consider this: by 2030, data centers are projected to consume as much as 12% of all electricity in the United States [1]. The International Energy Agency (IEA) projects that global electricity demand from data centers will more than double by 2030, reaching a staggering 945 terawatt-hours (TWh) [2]. That’s more than the current annual electricity consumption of many developed nations.
This is not some distant, abstract projection. The tech giants at the forefront of the AI revolution are already making massive investments to secure the power they need. Microsoft, Amazon, Google, and Meta are not just building data centers; they are building “digital foundries for silicon-based superintelligence,” and they are turning to nuclear power to keep the lights on. In a series of blockbuster deals, these companies are pouring billions into nuclear energy:
•Microsoft has struck a deal to power its data centers with the resurrected Three Mile Island Unit 1, now rebranded as the Crane Clean Energy Center [3].
•Amazon has invested $500 million in X-energy, a company developing small modular reactors (SMRs), with the goal of securing 5 GW of nuclear power [4].
•Google has committed to Kairos Power’s fleet of 500 MW SMRs [5].
These are not just business deals; they are a fundamental reshaping of our energy landscape, driven by the unchecked ambitions of a handful of corporations. The pursuit of AGI has become so paramount that the long-held concerns about nuclear safety are being swept aside in a torrent of investment and political lobbying. The question is no longer if we should build new nuclear plants, but how fast we can build them. And in this rush, the cautionary tales of the past are being dangerously ignored.
The Safety Shortcuts
In the original conversation that sparked this article, a powerful analogy was used: if you were taking a group of novices skydiving, you would triple- and quadruple-check their equipment. You would ensure every strap was secure, every parachute packed perfectly. When the stakes are that high, there is no room for error. And yet, as we leap into this new nuclear age, it appears we are not only failing to check our equipment, but actively dismantling the safety checks we once had in place.
The Trump administration’s executive orders in May 2025, which aim to quadruple U.S. nuclear capacity to 400 GW by 2050, have been accompanied by a push to “reform” the Nuclear Regulatory Commission (NRC) [6]. This “reform” is a thinly veiled attempt to fast-track approvals and cut “red tape,” but critics, including former NRC chairs, warn that it could have “unintended, dangerous consequences” [7]. The Union of Concerned Scientists has echoed these concerns, highlighting the risk of “corner-cutting” when the lines between promotion and regulation are blurred [8].
This is not just a hypothetical concern. The history of nuclear power is littered with examples of what happens when speed is prioritized over safety. The Vogtle nuclear power plant in Georgia, for example, was delayed by seven years and ran $30 billion over budget, a stark reminder of the complexities and risks involved in nuclear construction [9]. And while the new generation of Small Modular Reactors (SMRs) promises inherent safety features, the fact remains that none are yet fully commercialized in the U.S. We are, in effect, jumping out of the plane with a parachute that has never been tested in the real world.
The consequences of a nuclear accident are catastrophic, and the burden of that catastrophe would not be borne by the tech giants or the politicians who are pushing for this rapid expansion. It would be borne by the public, in the form of irradiated landscapes, displaced communities, and a legacy of environmental contamination that would last for generations. The skydiving analogy is apt, but perhaps incomplete. In this scenario, it is not just the novices who are at risk, but everyone on the ground below.
Winners and Losers
In this high-stakes gamble, the winners are clear. The tech titans—Microsoft, Amazon, Google, and Meta—stand to gain immense power and profit. By securing a dedicated, carbon-free energy source for their AI ambitions, they can continue their relentless pursuit of AGI without being constrained by the limitations of the public grid. The nuclear industry, long dormant, is experiencing a renaissance, with investors and uranium stockholders reaping the financial rewards of this new gold rush. And the government, eager to maintain a competitive edge in the global AI race, can claim a victory for “energy dominance” and technological leadership.
But for every winner, there is a loser. And in this story, the losers are the communities who bear the risks without reaping the rewards. The history of uranium mining in the United States is a story of environmental injustice, with Indigenous communities like the Navajo Nation suffering from devastating health consequences, including cancer and birth defects, from exposure to radiation [10]. The promise of jobs and economic development has often been a hollow one, leaving behind a legacy of contaminated land and broken promises.
The problem of nuclear waste also remains unsolved. The United States produces 2,000 metric tons of high-level nuclear waste every year, and there is still no long-term storage solution [11]. This waste, which will remain radioactive for thousands of years, is a ticking time bomb, a burden that we are passing on to future generations. The communities that are being asked to host this waste are often the same communities that have been marginalized and disenfranchised for decades, a stark example of environmental racism.
And what of the broader public? While the promise of cheap, clean energy is an alluring one, the reality is that the costs of this nuclear gamble are being socialized, while the profits are being privatized. If a nuclear accident were to occur, it would be the taxpayers who would foot the bill for the cleanup, the healthcare, and the long-term environmental monitoring. The tech companies, with their headquarters safely ensconced in Silicon Valley, would be insulated from the consequences of their own ambition.
The Superintelligence Timeline
The frantic pace of nuclear development is not just about powering our current AI models; it’s about preparing for the imminent arrival of Artificial General Intelligence (AGI) and, shortly after, Artificial Superintelligence (ASI). The timeline for this transition is alarmingly short. Many experts now believe that AGI could be achieved between 2025 and 2030, with ASI emerging within months or even weeks of that breakthrough. This is not a distant, science-fiction future; it is a reality that we are hurtling towards with little to no preparation.
The emergence of AGI will trigger an “intelligence explosion,” a recursive loop of self-improvement where the AI rapidly enhances its own capabilities, leaving human intelligence far behind. This exponential growth is what makes the timeline so compressed and so dangerous. Once an AGI can improve its own code, it can accelerate its own development at a rate that is impossible for humans to comprehend, let alone control.
This is where the nuclear gamble becomes truly terrifying. The same forces that are pushing for a rapid expansion of nuclear power are also racing to create a technology that could, in the worst-case scenario, become uncontrollable. The very power plants that are being built to fuel the development of AGI could one day be under the control of an ASI with goals and values that are completely alien to our own.
The conversation that inspired this article raised a crucial point: “in a battle between ethics and wealth, ethics doesn’t even stand a chance.” The pursuit of AGI is the ultimate prize in the tech world, a multi-trillion-dollar market that has the potential to reshape the global balance of power. In this high-stakes race, the temptation to cut corners on safety, both in the development of AI and in the construction of the power plants that fuel it, is immense.
Conclusion: A Choice of Futures
We are at a crossroads that will define the next century of human civilization. The choices we make in the coming months and years about nuclear energy and artificial intelligence will determine whether we create a future of unprecedented prosperity or unleash forces that spiral beyond our control. One path leads to a world where AGI and clean, abundant energy solve humanity’s greatest challenges. The other leads to a world where we have sacrificed safety and ethics on the altar of technological ambition, leaving the most vulnerable to bear the consequences of our reckless pursuit of power.
The conversation that began this exploration was not a condemnation of technology or progress. It was a plea for caution, for a more thoughtful and deliberate approach to the powerful forces we are unleashing. It was a reminder that the choices we make today will have consequences that will reverberate for generations to come.
We must demand a more transparent and inclusive conversation about the future of energy and AI. We must hold our leaders, both in government and in the private sector, accountable for the risks they are taking. We must insist that the pursuit of profit and power does not come at the expense of our safety, our environment, and our shared humanity.
The race to AGI is not a race that can be won by a single company or a single country. It is a race that we will either win together, or lose together. The choice is ours. Let us choose wisely.
References
[1] Data Center Energy Needs Could Upend Power Grids and Threaten the Climate. (2025, April 15). EESI. https://www.eesi.org/articles/view/data-center-energy-needs-are-upending-power-grids-and-threatening-the-climate
[2] AI is set to drive surging electricity demand from data centres, while offering the potential to transform how the energy sector works. (2025, April 10). IEA. https://www.iea.org/news/ai-is-set-to-drive-surging-electricity-demand-from-data-centres-while-offering-the-potential-to-transform-how-the-energy-sector-works
[3] President Trump Signs 4 Executive Orders to Deploy New… (2025, May 28). Holland & Knight. https://www.hklaw.com/en/insights/publications/2025/05/president-trump-signs-4-executive-orders
[4] Amazon, Google, Meta and Microsoft go nuclear. (2025, June 12). Trellis. https://trellis.net/article/amazon-google-meta-and-microsoft-go-nuclear/
[5] Big tech is gambling on nuclear. (2024, November 4). Public Enterprise. https://publicenterprise.org/big-tech-is-gambling-on-nuclear/
[6] President Trump Issues Sweeping Executive Orders… (2025, June 5). K&L Gates. https://www.klgates.com/President-Trump-Issues-Sweeping-Executive-Orders-Targeting-Nuclear-Regulation-6-5-2025
[7] Why the US must protect the independence of its nuclear… (2025, July 7). The Bulletin. https://thebulletin.org/2025/07/why-the-us-must-protect-the-independence-of-its-nuclear-regulator/
[8] New Executive Orders Undercut Nuclear Regulatory… (2025, May 28). NIRS. https://www.nirs.org/new-executive-orders-undercut-nuclear-regulatory-commission-authority-to-protect-public-safety-security/
[9] Vogtle Units 3 & 4. (n.d.). Georgia Power. Retrieved September 11, 2025, from https://www.georgiapower.com/company/energy-industry/energy-sources/nuclear/vogtle-3-4.html
[10] US Uranium Mining Legacy Still Harms the Navajo Nation. (2025, May 6). Union of Concerned Scientists. https://blog.ucs.org/chanese-forte/us-uranium-mining-legacy-still-harms-the-navajo-nation/
[11] Nuclear Waste and Environmental Justice. (n.d.). C-10 Research & Education Foundation. Retrieved September 11, 2025, from