Startups, billionaires and governments are betting big on next-gen reactors to fuel artificial intelligence's surging power needs.
Once left for dead after disasters like Fukushima and Chernobyl, nuclear power is making a dramatic return. But this time, it isn't fear of climate change or energy independence leading the charge. It's artificial intelligence. The surge in demand for electricity to run AI-powered data centres is turbocharging interest in small, modular nuclear reactors. Welcome to the new nuclear age, where the fuel isn't just uranium, it's ambition.
Next-gen nuclear for AI's growing appetite Artificial intelligence doesn't just run on code, it runs on colossal amounts of power. OpenAI alone has projected it will need 250 gigawatts by 2033; more than Brazil's entire consumption. Data centres already use around 40 GW, a figure expected to double within five years. Traditional energy sources are lagging. Natural gas is backordered, renewables remain intermittent, and coal is politically fraught. Enter small modular nuclear reactors (SMRs), now positioned as AI's most reliable partner in power.
Billionaires, believers, and backyard reactors
The money trail is explosive. Venture funding for new nuclear startups has skyrocketed from $500 million in 2020 to over $4 billion in 2025. Names like Sam Altman, Peter Thiel, Antonio Gracias and even Jeff Bezos are backing ventures such as Oklo, Aalo Atomics and X-energy. Their pitch is compelling: factory-made reactors that are smaller, cheaper, and faster to build than traditional plants. They promise 24/7 clean energy, often in a footprint smaller than a football pitch. For some, this race to power AI has led to a "code red" scenario for other AI development companies like OpenAI, as seen in the article, OpenAI CEO issues "code red" as Gemini hits 200M users.
Aalo Atomics' founder, Matt Loszak, is chasing a July 4, 2026 deadline set by Donald Trump to demonstrate reactor viability. His company plans to power data centres by 2027 using their prefabricated 10 MW reactors. Fellow trailblazers like Valar Atomics, founded by 26-year-old Isaiah Taylor, are targeting demonstration milestones within months. Their aim: to become the 'Toyota Corolla of nuclear': accessible, practical, and everywhere.
From military bases to gigafactories
The shift isn't just technological, it's political. Trump, since returning to office, has axed incentives for wind and solar while expanding nuclear tax credits up to 40%. He is fast-tracking permits, especially for reactors on military bases and historic nuclear sites like Idaho National Laboratory. It’s a strategy of pragmatism over purism: accelerate power deployment while bypassing local resistance.
This regulatory tailwind is spurring serious industrial ambition. Aalo is already eyeing a million-square-foot gigafactory. Fermi, the most lavishly funded of the lot, has secured 5,000 acres near a US Department of Energy site in Texas and ordered 2.5 GW of natural gas turbines while awaiting reactor approval. It plans to become the atomic epicentre for US AI. This kind of industrial expansion highlights the growing impact of AI on various economies, a topic also explored in AI & Robots Transform China's Economy.
The reactor renaissance is deeply personal
The people building this nuclear future often have intimate relationships with energy poverty or environmental degradation. Loszak credits his Canadian childhood asthma cure to nuclear replacing coal. Yasir Arafat, Aalo's CTO, studied nuclear engineering after doing homework by candlelight in Bangladesh. Their mission? Democratise clean, affordable energy worldwide, not just power Silicon Valley.
Others, like Kam Ghaffarian of X-energy and Mike Laufer of Kairos Power, have spent decades in aerospace or reactor development and are finally seeing mainstream traction. TRISO fuel, pebble-bed designs, molten salt cooling; these aren’t theoretical anymore. They are being fabricated, tested and fine-tuned in real facilities, from Utah to New Mexico. The rapid pace of development in this field is reminiscent of the discussions around AI's inner workings, which continue to baffle experts at major summits, as detailed in AI's inner workings baffle experts at major summit.
It’s not all green lights Regulatory delays, engineering setbacks, and public scepticism remain formidable. Oklo, despite being a stock market darling with a $15 billion valuation, may miss Trump's 2026 criticality deadline. Gates backed TerraPower, 17 years in the making, still awaits NRC approval. Lawsuits against the regulator argue for reform; even young startups like Deep Fission and Valar have joined the fray.
Then there’s the unpredictability of politics. Trump’s current embrace of nuclear may not last. But for now, the alignment of energy needs, financial capital, and executive muscle means the US and possibly Asia next is barreling into a nuclear revival not seen since the mid-20th century. For more on the regulatory landscape of nuclear power, the World Nuclear Association provides comprehensive insights into global policies and challenges World Nuclear Association.
Will Asia follow America’s lead and embrace small-scale nuclear to meet its AI demands? Or will regulatory caution and public resistance hold it back? The AI age is hungry, and only time will tell if atoms can satisfy its appetite. This question is particularly relevant when considering the broader impact of AI, including potential job losses, as discussed in AI's Job Impact: UK Faces Steep Employment Decline. Asia to Follow?.
Latest Comments (5)
Achcha, this is quite insightful. As someone following the tech landscape from India, I’ve often wondered about the sheer power consumption these advanced AI models demand. To think it's now directly fuelling a nuclear renaissance in the US is a fascinating pivot. It makes perfect sense, actually; traditional renewable sources, while brilliant, often struggle to provide the consistent, high-density power that data centres operating 24/7 require. This isn't just about clean energy, it's about *reliable* clean energy on a gargantuan scale. Small modular reactors seem like a very clever solution to bridge that gap. The article paints a compelling picture of necessity breeding innovation.
This is fascinating, especially the idea of AI driving a nuclear renaissance. I wonder, though, how these smaller reactors will address the long-standing *issue* of nuclear waste *disposal*? That's always been a sticking point with public *opinion*, no matter how much tech improves. Will these new designs offer a cleaner, more practical *solution*?
Blimey, AI's hunger for juice is bringing back the atom. Makes one wonder if this "renaissance" is really sustainable, or just a quick fix, innit?
"Wow, this is quite a read! It reminds me of the power outages we sometimes get here in the Philippines. You really feel the need for stable energy then. It’s fascinating how AI's demand could actually push for something as big as nuclear. Hopefully, the US gets it right this time; we’ve all heard those old stories, eh?"
Interesting read! While it's brilliant to see innovation, I can't help but wonder if the US's emphasis on SMRs might overlook the potential of established, larger nuclear plants, especially given the urgent energy needs. Are we sacrificing proven capacity for novelty, or is there a bigger picture I'm missing?
Leave a Comment