With the rise of AI creating an unprecedented thirst for energy, data centers are straining power grids and raising environmental alarms. To explore a potential solution, we’re joined by Christopher Hailstone, an expert in energy policy and grid reliability. We’ll discuss how small modular reactors, or SMRs, could offer a path to clean, reliable power for this booming industry, but only if they can overcome a web of Cold War-era regulations that have stifled nuclear innovation for decades. Our conversation will cover the unique fit between SMRs and data centers, the specific regulatory roadblocks hindering progress, the lessons learned from recent project failures, and the crucial reforms needed to unlock the future of nuclear energy.
AI data centers create an enormous and constant need for power. How are small modular reactors uniquely suited to meet this 24/7 demand compared to other energy sources, and what would the ideal on-site deployment model look like for a large tech company?
The synergy between SMRs and data centers is almost perfect. Data centers have this massive, concentrated, and, most importantly, constant energy load. They simply cannot tolerate intermittency. This is where SMRs shine, providing a steady, reliable stream of power around the clock. The ideal deployment model, which we’re seeing startups like Aalo Atomics pursue, is to build these compact, factory-built reactors on or near the data center campus itself. This co-location eliminates transmission losses and creates a dedicated, resilient power source. We’re talking about a demand of tens of gigawatts in the next five years alone—that’s like building 20 or 30 new cities. For a company like Meta or Google, having a dedicated SMR is the ultimate solution for both energy security and meeting their carbon-free goals.
The current nuclear licensing framework dates back to the 1950s and was designed for large-scale plants. What specific challenges does this one-size-fits-all system create for modern, factory-built SMRs, and how do these hurdles directly contribute to ballooning costs and project cancellations?
You’ve hit on the central problem. We’re trying to regulate a modern, modular technology with a framework finalized in 1956, designed for enormous, one-of-a-kind construction projects. This one-size-fits-all approach is incredibly inefficient. It subjects SMRs, which are designed for factory production and standardized designs, to the same fundamental licensing category as a giant, grid-scale plant. This means unnecessarily onerous costs and excruciatingly long wait times. These delays have a direct, devastating financial impact. Financing costs accumulate, inflation eats away at the budget, and schedule risk becomes a project-killer. This isn’t theoretical; we saw it happen in real-time with the Carbon Free Power Project.
Innovators often face a paradox where obtaining test data for a new nuclear design requires a license, but the license itself requires that data. From your perspective, how does this specific regulatory loop stifle development, and what are some practical steps to overcome it?
It’s a classic catch-22 that absolutely crushes innovation before it can even get off the ground. Developers are told they need to prove their design is safe with extensive test data, which is perfectly reasonable. However, the regulations make it, as one CEO put it, “impossible” to generate that data without first having the very license you’re trying to obtain. This regulatory loop creates an insurmountable barrier for startups and new designs. To overcome it, regulators need to adopt a more modern, phased, or milestone-based approach. This would allow innovators to conduct limited, supervised testing under a provisional permit to gather the necessary data, rather than demanding a complete data package for a final license from day one. It’s about creating a pathway for development, not just a wall.
The Carbon Free Power Project was cancelled the same year it received federal approval after years of delays and rising costs. What key lessons can new projects learn from this outcome, and what specific metrics should they monitor to avoid a similar fate?
The Carbon Free Power Project is a painful but powerful case study. The biggest lesson is that regulatory timelines are a primary driver of financial viability. The project was announced in 2015, but final certification didn’t come until 2023. In that eight-year span, the projected cost of power spiraled from roughly $60 to $90 per megawatt-hour. This increase was driven by financing costs, inflation, and the immense schedule risk created by the slow approval process. For new projects, the key metrics to monitor are not just construction costs, but the “cost of time.” They must rigorously track the burn rate associated with regulatory delays and build realistic, risk-adjusted financial models. When participating customers began to withdraw due to the ballooning costs, the project collapsed, teaching us that market confidence is directly tied to regulatory predictability.
A new regulatory framework, 10 CFR Part 53, is expected by 2027 to be better suited for advanced reactors. What are the most critical changes this new rule must contain to actually accelerate SMR deployment, and what risks might arise if it isn’t finalized soon enough?
The finalization of 10 CFR Part 53 is arguably the most critical event on the horizon for the SMR industry. For it to be successful, it absolutely must be what the NRC staff promised: a “risk-informed, technology-inclusive” framework. This means moving away from the prescriptive, one-size-fits-all model and instead evaluating new reactor designs based on their specific safety profiles and innovative features. It needs to create a clear, efficient, and predictable path to licensing. The risk if this rule is delayed or fails to deliver on its promise is enormous. We will see more project cancellations, and private capital will flee the sector. With the demand from data centers growing exponentially, a delay past 2027 means we miss this crucial window of opportunity, and that demand will be met by fossil fuels, pushing U.S. greenhouse gas emissions even higher.
What is your forecast for the SMR industry’s ability to meet the urgent energy demands of AI infrastructure over the next decade?
My forecast is one of cautious but determined optimism. The demand is undeniable, and the private sector interest is incredibly strong—tech giants are actively seeking out SMRs and are even willing to pay a premium for that reliable, clean power. The technology is advancing rapidly. The entire equation, however, hinges on regulatory reform. If 10 CFR Part 53 is implemented effectively and on time, and we successfully cut through the remaining red tape, I believe we will see the first SMRs powering data centers within the next decade. But if the regulatory logjam persists, the industry will struggle to get projects built fast enough, and a historic opportunity to align our digital and climate goals will be tragically missed.
