Today we’re speaking with Matilda Bailey, a networking specialist whose work puts her at the forefront of the technologies and infrastructure powering our digital world. As AI workloads create unprecedented energy demands, the conversation has turned to advanced nuclear power, specifically Small Modular Reactors, or SMRs. A recent regulatory challenge to a proposed SMR project in Texas has brought the promises and perils of this technology into sharp focus, making this a critical moment to explore the intersection of data center needs, energy innovation, and policy.
Our discussion will delve into the complex realities facing SMR adoption. We’ll explore the financial and regulatory hurdles highlighted by the Texas case, contrast the unique operational advantages of SMRs against their significant challenges, and outline a practical roadmap for data center operators navigating this uncertain landscape. We’ll also touch on the importance of community engagement, the potential of hybrid energy systems, and what the future might hold for nuclear-powered data infrastructure.
A recent regulatory decision in Texas allowed an intervention against a proposed SMR project based on financial qualifications. What specific financial and scheduling risks does this signal for future projects, and how can operators proactively build a more defensible commercial structure to mitigate these hurdles?
That decision in Texas was a real wake-up call, and it’s something I’ve been discussing a lot. While it doesn’t mean the project is dead in the water, it absolutely signals that the path forward will be under a microscope. The immediate risks are significant delays to the schedule and, of course, ballooning financial costs as legal and regulatory processes drag on. It’s a clear message that technical approval isn’t the only barrier; the commercial viability and financial backing of these projects will be just as fiercely scrutinized. To get ahead of this, operators need to stop treating SMRs like a typical power purchase and start treating them like the major infrastructure projects they are. This means building commercial structures with concrete milestones, clear off-ramps if those milestones aren’t met, and a transparent allocation of risk between all partners. You can’t just assume licensing will be a routine check-box; you have to build a business case that can withstand intense public and regulatory pressure from day one.
Given the step-function increases in energy demand from AI, some view nuclear as a “strategic hedge.” Could you walk through the key operational advantages SMRs offer over renewables, such as land use and baseload power, and explain how these benefits compare to their significant regulatory challenges?
The term “strategic hedge” is perfect because it captures the core dilemma. The demand from AI isn’t growing incrementally; it’s exploding in massive, step-function increases that the grid simply wasn’t built for. This is where SMRs present a compelling case. First, they provide 24/7 firm, carbon-free power. Unlike solar and wind, which are intermittent, an SMR delivers consistent baseload energy without the need for massive, expensive battery storage systems. Second is the physical footprint. SMRs require significantly less land than a solar or wind farm of comparable output, which is a huge advantage in many regions. And because they’re modular, you can scale your power generation to match demand growth, potentially deploying a 50 MW unit now and adding more later, avoiding overbuilding. These advantages are immense, but they come face-to-face with the reality of nuclear regulation. The process is inherently stricter, the timelines are longer, and as we’ve seen, the points of failure aren’t just technical—they’re financial, political, and social. So, the strategic bet is whether these powerful operational benefits are worth the grueling, and often unpredictable, journey through the regulatory landscape.
The U.S. is developing a new regulatory framework, Part 53, for advanced reactors, which isn’t expected until 2026. What specific steps should data center operators take in the meantime to navigate this uncertainty and build trust with regulators, communities, and planning authorities for proposed SMR projects?
The fact that Part 53 is still years away—expected around March 2026—creates a fog of uncertainty that can be paralyzing. But operators can’t afford to just wait. The most critical action is to engage, and engage early. Don’t wait until you have a perfect plan. Start building relationships with regulators, local planning authorities, and community leaders now. Success in this space hinges as much on trust and consent as it does on engineering. You need to be on the ground, using plain, accessible language to explain the safety features, the environmental benefits, and the resilience your project offers. If you lose the confidence of any of these groups, the project can slow to a crawl or stop entirely, regardless of what the final regulations look like. Think of this period not as a waiting game, but as a foundational period for building the social and political license to operate. Regulatory friction should be treated as a permanent, structural feature of this landscape, not a temporary problem.
You’ve highlighted the need to address difficult topics like waste management and water use early on. Can you describe a step-by-step process for how a project team should approach these conversations with local stakeholders to build confidence and secure consent from the very beginning?
This is where many projects stumble—by shying away from the tough questions. The key is to be proactive, not reactive. The first step is to do your homework and map out every single one of these “hard topics” before you even think about a public meeting. This means waste management, emergency planning, water usage, local environmental impacts, and how the facility will interact with the existing grid. Step two is to initiate dialogue with small, trusted community groups and leaders, not a massive town hall. This allows for genuine conversation, not grandstanding. In these meetings, you lead with those difficult topics. You say, “We know you have questions about waste and water, and we want to address those first.” Step three is to present your plans with transparency and humility, acknowledging what you know, what you don’t, and how you plan to find the answers. Finally, step four is to create permanent channels for feedback and incorporate that feedback visibly into your plans. This isn’t about “selling” a project; it’s about co-creating a solution and showing the community that their concerns are being engineered into the project’s DNA from the very start.
Experts recommend designing holistic power systems that combine SMRs with renewables and storage. What would a hybrid model like this look like in practice for a large data center campus, and how does this approach help make a project more defensible to both regulators and investors?
A hybrid model is an incredibly powerful strategy because it demonstrates a sophisticated, resilient approach to energy. For a large data center campus, it would likely involve an SMR, say 100 or 200 MW, providing the constant, reliable baseload power that AI infrastructure absolutely must have. This is your bedrock. Layered on top of that, you would have on-site or nearby solar generation and a battery storage system. During the day, the solar panels help meet peak demand and charge the batteries. During SMR maintenance outages or unexpected grid events, the batteries provide an immediate bridge, ensuring uninterrupted service. This holistic design is far more defensible. To regulators, it shows you’re not just betting on one technology but are building a resilient system that integrates multiple carbon-free sources and addresses reliability head-on. To investors, it de-risks the project; it’s no longer an all-or-nothing nuclear play but a diversified energy portfolio tailored to the specific needs of the mission-critical facility.
The inflection point for SMR adoption is described as when they have predictable timelines and bankable financing. Based on current progress, what specific milestones or metrics should we watch for over the next five years to know if SMRs are successfully making that transition?
That’s the billion-dollar question. To know if we’re truly at that inflection point, I’d watch for three key milestones over the next five years. First, we need to see the first handful of SMR projects move from licensing application to construction on a predictable, repeatable schedule—not one-off science projects, but a real process. If we see a design get through the NRC and break ground within a consistent 36-to-48-month window multiple times, that’s a huge signal. Second, we need to see major financial institutions—not just venture capital or government grants—providing traditional project financing for an SMR. When a project becomes “bankable,” it means the risk profile is understood and accepted by the broader market. Finally, I’d look for the emergence of standardized commercial contracts and insurance products specifically for SMRs. When you can buy off-the-shelf insurance for a project, it means the risks have been quantified and the industry is maturing from pilot projects to a proven, commercial infrastructure class.
What is your forecast for small modular reactors powering data centers?
My forecast is one of cautious but determined optimism. In the short term, the next five to seven years will be fraught with challenges. We will see more regulatory hurdles like the one in Texas, and some early projects will likely face delays or even cancellations. However, the fundamental driver—the insatiable energy demand of AI—is not going away. Power is rapidly becoming the primary limiting factor in AI expansion. Because of this, the strategic necessity of a firm, carbon-free, scalable power source will force the industry to solve these regulatory and financial puzzles. I predict that by the early 2030s, we will see the first operational SMRs powering major data center campuses. It won’t be a floodgate opening all at once, but a steady, deliberate deployment at key strategic sites, paving the way for broader adoption as the technology proves itself not just as an engineering marvel, but as a reliable and bankable asset.
