Last Energy plans to deploy 30 microreactors in Texas to power AI data centers, tackling the growing energy crisis.
The breakneck expansion of AI has pushed data centers to an unsustainable energy consumption trajectory in the United States. As enterprises race to deploy ever-larger AI models, the power demands of hyperscale data centers have surged. In 2023, data centers accounted for 4.4% of total U.S. electricity consumption, a figure that could rise to between 6.7% and 12% by 2028, according to a recent DOE report.
Lenovo AI Servers
With AI’s computational intensity skyrocketing, Goldman Sachs estimates that data center electricity usage could double by 2030. For instance, Google’s AI infrastructure now consumes enough power to rival entire mid-sized cities, prompting the company to explore nuclear solutions.
Last Energy’s Microreactors: A Scalable Power Solution for AI Data Centers
As the power grid struggles to keep pace with AI-driven demand, Last Energy has announced plans to deploy 30 microreactors in Texas specifically to power data centers (Last Energy). The reactors will be built on a 200-acre site in Haskell County, Texas, with Last Energy filing for grid connection approval from ERCOT. The company is also preparing an Early Site Permit application with the U.S. Nuclear Regulatory Commission, a critical step toward commercial deployment.
Rendering of Last Energy’s Northwest Texas Project
Last Energy’s PWR-20 microreactors (20 MWe) are designed for modular manufacturing, meaning they can be built off-site and rapidly deployed to meet power needs. The company has built two full-scale prototypes in Texas and demonstrated its design to industry and government officials.
“Texas is America’s undisputed energy leader, but skyrocketing population growth and data center development is forcing policymakers, customers, and energy providers to embrace new technologies,” said Bret Kugelmass, Founder and CEO of Last Energy. “Nuclear power is the most effective way to meet Texas’ demand, but our solution — plug-and-play microreactors, designed for scalability and siting flexibility — is the best way to meet it quickly. Texas is a state that recognizes energy is a precondition for prosperity, and Last Energy is excited to contribute to that mission.”
Why Microreactors? A Long-Term Solution for AI Infrastructure
Unlike conventional large nuclear plants, microreactors offer a decentralized, flexible, and zero-emission power source, making them ideal for AI and high-performance computing (HPC) environments. The U.S. government and private companies have increasingly turned to microreactors to ensure energy resilience, with projects like:
- TerraPower & Sabey Data Centers: Exploring using Natrium microreactors to power data centers in Texas and the Rocky Mountain region (DataCenter Dynamics).
- Google’s Nuclear Investments: Partnering with Oklo and Holtec to integrate small modular reactors (SMRs) into data center infrastructure (AP News).
Microreactors offer a compelling solution for AI-driven power growth because they:
- Generate consistent, high-density power without requiring vast land or fuel logistics.
- Reduce carbon emissions, making them attractive for ESG and sustainability initiatives.
- Bypass traditional grid limitations, ensuring power availability even in constrained regions.
The Future: Can Microreactors Reshape AI Infrastructure?
The traditional energy grid will struggle to sustain data center expansion with AI models consuming increasingly demanding power levels. Microreactors and nuclear energy, generally, could shift the balance, offering dedicated, low-carbon, high-efficiency power tailored for hyperscale computing.
If Last Energy’s 30-reactor deployment in Texas proves successful, similar micro-nuclear projects could emerge across Europe, Silicon Valley, Northern Virginia, and other AI hotspots, where power grid limitations already pose a risk to future data center growth.
Engage with StorageReview
Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | TikTok | RSS Feed
