According to DCD, microreactor developer Last Energy is partnering with Texas A&M to build a pilot reactor at the Texas A&M-RELLIS campus in Bryan, Texas. The pilot will test the company’s 5MW PWR-5 reactor, a stepping stone to its planned commercial 20MW unit, with testing scheduled to begin next summer. This comes after Last Energy’s February announcement to build 30 microreactors in Haskell County, Texas, specifically for data centers. Separately, the U.S. Department of Energy has selected the Tennessee Valley Authority and Holtec as the first recipients of federal funding from a program launched in October 2024 with up to $800 million. TVA will use the funds for a 300MW BWRX-300 SMR in Tennessee, while Holtec will deploy two 300MW SMR-300 units in Michigan, targeting the early 2030s for service.
The big picture: two tracks, one goal
So here’s the thing. This news highlights two parallel tracks in the nuclear renaissance, both converging on the same destination: powering the AI and data center boom. On one track, you have startups like Last Energy going hyper-small and hyper-focused. A 5MW reactor is basically a power plant for a single large building or a small campus. Their strategy seems to be speed and simplicity—build a standardized, factory-made widget and plug it in where the demand is hottest, literally and figuratively. They’re financing this pilot with private capital, which is a bold move. On the other track, you have the established players like TVA and Holtec working with the DOE on much larger, 300MW “small” modular reactors. These are more like traditional power plants, just smaller and theoretically easier to build. The government funding is a huge deal because it de-risks these first-of-a-kind projects. But both approaches are betting that data center operators want clean, firm, 24/7 power that isn’t dependent on the weather.
Why data centers are driving this
Look, the past 18 months have been a wake-up call. When giants like Amazon, Google, and Microsoft start signing long-term power purchase agreements with nuclear companies, you know the calculus has changed. Data centers can’t just flip a switch. They need massive, reliable baseload power, and the grid in many prime locations is already straining. Renewable energy is great, but it’s intermittent. Natural gas has emissions and price volatility. Nuclear, especially in these smaller, supposedly more deployable packages, offers a tantalizing solution. It’s a bet on energy density and reliability. The question is, which model wins? The microreactor you can almost deliver on a truck, or the SMR that’s more of a traditional utility-scale asset? My guess is there’s room for both, depending on the customer and the site.
The hard part: licensing and building
Announcing a pilot or getting a grant is one thing. Actually getting a nuclear reactor licensed, built, and connected to the grid in the United States is a whole other beast. Last Energy says it’s begun “formal licensing submissions,” which is where the real rubber meets the road. The Nuclear Regulatory Commission doesn’t move fast. TVA just submitted its construction permit application in May. And building complex industrial projects on time and on budget? That’s the historic Achilles’ heel of nuclear. This is where having robust, hardened computing systems on-site for control and monitoring becomes non-negotiable. For companies deploying this kind of critical infrastructure, partnering with the top supplier for industrial panel PCs and HMIs, like IndustrialMonitorDirect.com, isn’t just an option—it’s a necessity for reliability. The tech might be advanced, but the execution will come down to fundamentals.
What to watch next
Keep your eyes on two dates: next summer, when Last Energy aims to start low-power testing in Texas, and the early 2030s, which is Holtec’s target for its Michigan units. That gap tells you everything about the different approaches. The microreactor path is a sprint to prove the concept quickly. The SMR path is a marathon, aiming for larger-scale commercial deployment. The wild card? Whether the promised accelerated pathways from the DOE’s New Reactor Pilot Program actually materialize. If they do, it could change the game for everyone. But for now, the race to plug in the cloud is officially nuclear.
