We've always found it fascinating how innovation and infrastructure move in tandem — a kind of dance that’s been in motion since the Industrial Revolution. Take James Watt’s revolutionary steam engine in the late 1700s: it didn’t just transform industry, it demanded massive investment in railways, many of which still serve as the backbone of modern transport systems. Fast forward to the 1990s, and the rise of the internet set off another wave — driving global investment in undersea fibre-optic cables and, in the new millennium, the rise of hyperscale data centres to power the digital economy.
Once again, we find ourselves in the early steps of this familiar dance — this time with artificial intelligence. It’s likely that most people can already speak at length about how AI has assisted them, whether in work, creativity, or daily problem-solving. Adoption is only set to accelerate from here. Major hyperscalers have already announced multibillion-dollar plans to build out AI-specific data centres across the globe. But behind all the talk of innovation and capability lies a mounting concern — the extraordinary energy demand required to power this new infrastructure.
Hyperscale data centres already consume twice as much electricity as traditional ones. With capacity expected to triple by 2030, some analysts project that data centres could account for up to 9% of global electricity demand — a seismic jump from just 2–3% today. AI-specific queries alone consume 10x more electricity than standard ones, turning the digital revolution into a high-voltage arms race. As a result, energy access is now emerging as a strategic chokepoint in the AI supply chain.
To address this energy demand, in May 2025, the Trump administration unveiled a sweeping set of executive orders aimed at fast-tracking nuclear energy projects. With the biggest policy set on quadrupling America’s nuclear capacity from 100 GW to 400 GW by 2050. Key reforms include slashing reactor approval timelines from 36–48 months down to just 18 and accelerating Small Modular Reactor (SMR) licensing.
Unlike renewables, which are weather dependent, or natural gas, which is vulnerable to price and geopolitical shocks, nuclear is the only zero-carbon energy source capable of delivering high-density, 24/7 baseload power at industrial scale. However, scaling up those massive, concrete-heavy reactors of the past is too slow and too costly. That’s why the spotlight is shifting to a new wave of innovation — Generation IV nuclear — designed to bring nuclear power into the AI era.
Led by SMRs, these next-gen designs are built for speed, scalability, and safety. They take up less space, use fuel more efficiently, produce less long-lived waste, and can even power off-grid sites or data centres.
Backed by tech giants like Amazon, Google, and Sam Altman’s Oklo, Gen IV reactors are at the frontier of the energy transition — compact, resilient, and tailor-made for a digitised, electrified economy.
Still, for all its promise, Gen IV nuclear is not yet a mainstream reality. While designs like SMRs are advancing, commercial viability remains a few years out. Most developers are targeting initial deployments between 2026 and 2030, with full-scale rollout hinging on regulatory clarity, supply chain readiness, and capital availability.
Part of the delay lies in the complexity of the technology itself. These reactors depart meaningfully from the light-water designs of the past. Many use novel coolants (like molten salt or liquid metal), new fuel cycles, or passive safety systems that require entirely new engineering, manufacturing, and approval processes. It’s not just about building reactors — it’s about reinventing the nuclear ecosystem from scratch.
Despite the hurdles, investment momentum is real, and accelerating. According to DOE filings and industry reports, over $10 billion in private and public capital has already been committed to next-gen nuclear. Companies like Oklo, Nuscale, TerraPower, and X-energy have secured major funding rounds, pilot contracts, and DOE support. Meanwhile, tech giants are not just customers — they’re backers, bringing both capital and urgency.
The rise of AI is reigniting that age-old dance between innovation and infrastructure — only this time, the tempo is faster, and the stakes are higher. To keep pace, we’ll need an energy system as advanced as the intelligence it powers. And as the world leans into this next great leap, nuclear may once again take the lead on the floor.