Powering the AI Revolution

Finance is just one example. Nearly every industry, domain, and knowledge function is being re-imagined with AI. In a recent interview with Fareed Zakaria on CNN’s Global Public Square, LinkedIn co-founder and AI venture capitalist Reid Hoffman called it the Cognitive Industrial Revolution, citing parallels to the dramatic innovations of the Industrial Revolution in the 19th Century.

Just as the steam engine powered the remarkable inventions of that fervent era, today’s AI-driven operations are accelerating the search for stable, carbon-free energy solutions to sustain today’s burgeoning demands. Goldman Sachs Research estimates that AI could drive a 160% increase in data center power demand by 2030, translating to an additional 200 terawatt-hours per year of electricity consumption.

Economics of Small Modular Reactors

According to a recent report by S&P Global Market Intelligence, despite the obvious potential of SMRs, the financial viability of these emerging power plants remains uncertain, with commercial-scale deployment still 10 to 15 years away. However, tax incentives and government support may help bridge the cost gap, making SMRs an attractive option for large-scale energy consumers such as data centers. Current projections based on estimates from the U.S. Energy Information Administration suggest that SMRs entering service in 2035 could produce electricity at a first-year power purchase agreement (PPA) price of approximately $79 per megawatt-hour.

While this cost is higher than some existing energy sources, it remains within an acceptable range for hyperscale data center investors seeking reliable, carbon-free power—especially given the sustainability objectives of these companies.

According to the S&P Global Market Intelligence report, Amazon has invested $500 million in X-Energy, with plans to deploy up to five gigawatts of SMR capacity. Google has signed an agreement with Kairos Power LLC to purchase 500 megawatts of energy from multiple SMRs, set to come online between 2030 and 2035. And Meta has announced a request for proposals to identify nuclear energy developers, targeting one to four gigawatts of new nuclear generation capacity—to cite three prominent examples. “Big Tech’s aggressive pursuit of nuclear energy solutions highlights a shift in how hyperscalers view their energy needs,” the report states. This aligns with the broader trend of technology companies seeking to lock in long-term, stable power sources that can complement intermittent renewables and support their massive data processing requirements.

In some cases, these new nuclear facilities will be placed near existing nuclear facilities, where there is already mature power infrastructure in place to facilitate an easy integration into the electrical grid. But will these new power plants be ready in time? IDC, a market research firm based in Needham, Mass., expects global data center electricity consumption to more than double between 2023 and 2028 with a five-year CAGR of 19.5%, reaching 857 terawatt hours in 2028.

The lengthy build time for SMRs presents a significant challenge in meeting these escalating energy needs. While the average construction timeline for a new data center ranges from 18 to 24 months—sometimes even faster—SMRs require an estimated six years to become operational. This timeline—which is fairly optimistic given the history of nuclear development in the U.S.—will make it difficult for SMRs to serve as an immediate solution to the rapidly growing energy demands of hyperscale technology companies. In contrast, alternative energy sources, such as gas combined-cycle plants, take an average of three years to build, while hybrid solar-plus-storage projects can be completed in just two years.

Despite this discrepancy in build times, nuclear energy remains attractive due to its unmatched reliability and efficiency. Once operational, nuclear reactors boast the highest capacity factor of any energy source—93.1% in 2023—compared to 33.2% for wind and 23.5% for solar. This performance advantage suggests that, although SMRs cannot address immediate energy shortages, they could play a crucial role in securing long-term, stable energy supplies for industries requiring consistent power.

The Security and Energy Independence

While SMRs offer long-term stability and efficiency, their viability is also intertwined with broader geopolitical considerations. Chief among these is the security of the nuclear fuel supply—a factor that directly impacts the resilience of the U.S. nuclear sector. In 2023, 95.4% of the uranium used in U.S. commercial reactors was imported, and four countries accounted for more than 82% of the total supply: Canada, Kazakhstan, Australia, and Russia. This heavy dependence on foreign uranium could jeopardize the long-term viability of the U.S. nuclear sector, especially given shifting international relations and geopolitical tensions. Russian uranium made up 12% of imports in 2023. Domestic uranium production remains low, making up less than 5% of U.S. civilian reactor needs. However, recent policy shifts, including the Inflation Reduction Act and renewed interest in domestic uranium mining and processing, suggest a push toward strengthening the U.S. nuclear fuel supply chain.

Securing a reliable nuclear fuel supply is not just a matter of energy independence—it also plays a crucial role in supporting industries that rely on stable, high-capacity power. Nowhere is this more evident than in the rapidly expanding data center sector, where power consumption is reaching unprecedented levels. Infrastructure providers such as Hitachi Vantara, which deliver compute and storage solutions for hyperscale data centers, are directly impacted by these shifts.

As AI workloads become more complex and energy-hungry, ensuring uninterrupted, sustainable power sources becomes critical for maintaining operational efficiency and meeting service-level agreements. Without stable power, even the most advanced computing and data storage architectures face performance bottlenecks and potential downtime, affecting industries from financial services to manufacturing to healthcare.

A Fundamental Shift

The race to secure the future of AI-driven infrastructure is well underway, and nuclear power is poised to play a pivotal role. The growing need for reliable, carbon-free energy isn’t just a challenge—it’s an opportunity for industries to rethink their power strategies, build resilience, and invest in long-term sustainability. SMRs represent a turning point in energy innovation, offering a way to bridge the gap between today’s escalating demand and tomorrow’s cleaner, more efficient power grid.

For companies that depend on data centers, the stakes are high: ensuring seamless operations while keeping pace with increasingly strict ESG mandates. SMRs present a compelling solution—delivering reliable, carbon-free energy at scale. Businesses that embrace resilient, sustainable power sources today won’t just meet regulatory expectations; they’ll position themselves as industry leaders in a rapidly evolving digital landscape. As nuclear energy, hyperscale infrastructure, and enterprise technology converge, forward-thinking investments aren’t just beneficial—they’re essential for long-term success.

Deployment challenges remain, but the momentum behind these next-generation nuclear solutions signals a fundamental shift in how we power the digital economy. For data centers, infrastructure providers, and corporations alike, the question is no longer if nuclear energy will be part of the solution—but how quickly the industry can adapt and integrate these nimble solutions into their long-term energy strategies.