The growth of artificial intelligence (AI) data centers is accelerating electricity demand across the United States, fueling interest in large-scale nuclear power projects capable of supplying reliable, carbon-free energy. However, a new report from the Institute for Energy Economics and Financial Analysis (IEEFA) warns that Fermi America’s proposed 4,456 MW nuclear campus could face significant engineering, regulatory and financial obstacles that may delay construction and inflate costs.
Fermi America plans to build four Westinghouse AP1000 reactors near Amarillo, Texas, supporting the U.S. administration’s objective of expanding nuclear generating capacity by 100 GW to meet rising demand from AI computing, cloud infrastructure, advanced manufacturing and industrial electrification. While the AP1000 is an established reactor design, IEEFA notes that Fermi America is only one year old and has yet to complete a commercial power generation project.
Unproven Cooling Technology Raises NRC Approval Risks
A concern highlighted by IEEFA Energy Finance Analyst Dennis Wamsted is Fermi America’s decision to replace the AP1000’s conventional water-based cooling system with an Air-Cooled Condenser (ACC) or hybrid cooling design. No large commercial nuclear reactor in the United States has ever operated using anything other than liquid-water cooling, making the proposed design a first-of-its-kind project requiring additional review by the U.S. Nuclear Regulatory Commission (NRC).
The IEEFA report argues that changing a standardized reactor design removes many of the licensing, construction and cost advantages that the nuclear industry has spent years trying to achieve.
Regulatory uncertainty has increased further because Fermi America’s NRC filings reportedly left critical engineering information incomplete, with sections describing the cooling system carrying the placeholder statement “TO BE PROVIDED LATER.” According to the report, unresolved engineering details during licensing increase uncertainty for regulators, investors and prospective customers while raising the risk of project delays and multibillion-dollar cost overruns.
Water Constraints Drive Alternative Cooling Strategy
Fermi America’s site in the water-stressed Texas Panhandle explains its search for alternative cooling technologies. The report notes that the two operating AP1000 reactors at Vogtle Units 3 and 4 are licensed to use up to 43,000 gallons of water per minute, equivalent to approximately 22.6 billion gallons annually or 69,449 acre-feet.
However, IEEFA cites previous studies concluding that air cooling is not technically suitable for large AP1000 reactors. Southern Company determined in 2009 that air-cooled condensers were not technically feasible because they require significantly larger equipment, reduce plant efficiency and substantially increase capital costs.
Cooling Costs Could Reach Hundreds of Millions
According to the report, previous estimates placed capital costs for air-cooled systems between $285 million and $445 million, increasing to approximately $428 million-$668 million in current dollars.
Hybrid cooling systems were estimated at $206 million-$324 million in 2013, but updated calculations suggest costs now exceed $300 million and could approach $500 million for systems designed to maximize water savings.
Operating expenses also increase significantly. EPRI estimated annual operating and power penalty costs of $23.4 million-$26.1 million for hybrid cooling compared with $9.2 million for conventional wet cooling. Installation costs for hybrid systems could be 6-9 times higher than traditional water-cooled designs.
Financial Challenges Add Pressure
IEEFA also points to growing financial concerns surrounding Fermi America. Since its public market debut in October, the company’s stock has declined by more than two-thirds from its opening price and has traded below $10 since early March.
The company has also experienced leadership turmoil, with co-founder Toby Neugebauer removed as chief executive officer and chief financial officer Miles Everson resigning. The report notes that the project’s estimated cost has reached approximately $41 billion, while the expected startup date for the first reactor has slipped from 2032 to 2033.
Rivals Advance Standardized Nuclear Designs
While Fermi America pursues a customized large-reactor strategy, several competitors continue advancing standardized nuclear technologies.
Oklo holds approximately $2.5 billion in cash with no outstanding debt and is developing factory-built Aurora Powerhouses ranging from 15 MW to 75 MW.
NuScale Power has approximately $1 billion in available liquidity and is advancing a 6 GW project pipeline through partnerships with the Tennessee Valley Authority and ENTRA1, requiring production of 72 standardized reactor modules.
Meanwhile, Valar Atomics has achieved first criticality for its Ward 250 gas-cooled microreactor designed for AI computing infrastructure, while Aalo Atomics has completed first criticality of its Aalo-X test reactor at Idaho National Laboratory to support commercialization of standardized 10 MWe Aalo Pods.
Standardization Remains the Industry’s Biggest Advantage
IEEFA concludes that standardized reactor designs remain the nuclear industry’s strongest tool for reducing construction risk, lowering capital costs and accelerating regulatory approvals. Although AI-driven electricity demand is creating a major growth opportunity for nuclear energy, the report warns that Fermi America’s decision to modify the licensed AP1000 design introduces substantial technical, regulatory and financial uncertainty.
Unless the company successfully addresses these engineering challenges, IEEFA believes the project could face higher costs, longer construction schedules and increased financing risks, while competitors deploying standardized reactor platforms continue strengthening their position in the emerging AI-powered energy market.
SHAFANA FAZAL
