Fermi America has won preliminary regulatory support for one of the largest proposed private-grid nuclear projects in the United States. Staff at the U.S. Nuclear Regulatory Commission have recommended issuing combined licenses for four Westinghouse AP1000 reactors at Project Matador near Amarillo, Texas.
The draft environmental impact statement, released for public comment on October 9, covers a project that would add more than 4.4 gigawatts of nuclear generation capacity. For investors, the decision matters because it moves Fermi’s Texas reactors beyond concept stage, even as the company still faces financing, safety review and execution risk.
The milestone also arrives as Fermi tries to stabilize its business after management upheaval, a failed funding arrangement and delays in bringing initial on-site power online for data-center customers.
Key Facts
- NRC staff preliminarily supported combined licenses for four Westinghouse AP1000 reactors at Project Matador on October 9.
- The four-unit project would add more than 4.4 GW of nuclear generation capacity to Fermi’s planned private grid near Amarillo, Texas.
- Fermi’s environmental review reached draft stage roughly 15 months after application under an NRC pilot allowing applicants to prepare their own review documents.
- Fermi signed a binding 15-year lease in August with TensorWave for 222 MW valued at about $6.5 billion, with options to expand beyond 650 MW.
- Fermi is targeting first reactor operations by 2032, while the final environmental review is expected in December and the safety review remains unscheduled.
Fermi Texas Reactors
The draft environmental finding is a meaningful step for Fermi Texas reactors because it reduces one layer of regulatory uncertainty for a project designed to pair large-scale power generation with artificial intelligence infrastructure. Project Matador is intended as a combined energy and computing campus, reflecting a wider market push to secure dedicated power for data centers as AI workloads expand.
Still, the environmental review is only one part of the licensing path. The more consequential hurdle may be the safety review, which does not yet have a published timeline. Fermi also withdrew the site-specific portion of its application and plans to refile it by December 31, after which a mandatory hearing and a commission vote would follow. That means the path from preliminary support to final approval remains long.
The timing matters because Fermi’s commercial strategy depends on convincing tenants and financiers that power will arrive on schedule. The company has already revised expectations for early electricity supply. In late 2025 it had projected 1 GW of capacity online by late 2026, but by August 2026 that outlook had shifted to about 200 MW of initial commercial power over the next six months. For prospective customers, that change underscores how difficult it is to build power infrastructure and compute campuses simultaneously.
Preliminary environmental backing is an important milestone, but Fermi’s real test is converting regulatory momentum into financed, contracted and deliverable power.
Why the tenant and financing picture matters
Fermi’s commercial narrative improved in August when it signed a binding 15-year lease with TensorWave, an AMD-focused GPU cloud provider. The contract covers 222 MW and carries an estimated value of about $6.5 billion, with expansion options that could take demand above 650 MW. Phased delivery is scheduled to begin in the second half of 2027.
Even so, the lease has not yet closed. The deadline was extended from September 30 to October 31 to complete a lease guarantee, and project financing is still needed. AMD’s agreement to provide credit support for the lease adds credibility, but investors may also scrutinize the interconnected nature of the arrangement: AMD supplies TensorWave’s chips, has invested in TensorWave and is now backing its lease obligations. That structure may help close the deal, yet it also raises questions about concentration risk and the durability of end demand.
On the operating side, Fermi has assembled significant thermal generation equipment for near-term power. The company has taken delivery of six Siemens SGT-800 turbines, three Siemens F-class units and three GE Vernova 6B turbines. NAES has been selected to operate the fleet, while Hillcore Energy Capital signed a framework agreement to finance, build and own roughly 2.6 GW of additional on-site capacity. Those steps suggest Fermi is trying to bridge the gap between immediate data-center power needs and the much longer development cycle for nuclear generation.
Implications for Investors
For investors tracking nuclear energy, grid infrastructure and AI-related power demand, Fermi’s progress offers both opportunity and caution. The opportunity lies in the scale of the proposed project: more than 4.4 GW of planned nuclear capacity tied to a fast-growing market for dedicated data-center power. If Fermi can secure approvals, financing and anchor tenants, Project Matador could become a notable example of AI demand reshaping U.S. generation development.
The risks are substantial. Licensing is incomplete, the safety review timeline is unknown and site-specific filings still need to be resubmitted. Commercial execution also remains under pressure. A prior prospective tenant withdrew from a $150 million construction funding agreement, triggering a sharp selloff in the company’s shares. Governance has been another overhang after the April removal of co-founder and chief executive Toby Neugebauer, the departure of the chief financial officer and ongoing legal disputes involving board actions and shareholder rights.
Leadership changes may help reset the story. In August, the board appointed Lee McIntire, a former TerraPower chief executive and Bechtel veteran, as CEO. Investors will now be watching for three concrete signals: closure of the TensorWave lease, evidence that near-term on-site power can be delivered close to the updated schedule, and a clearer NRC roadmap for the safety review and final licensing process.
The next phase for Fermi will determine whether this NRC milestone becomes a foundation for project financing or simply another interim step in a complex buildout. For the broader market, the case is worth watching as a real-world test of whether nuclear power can be developed fast enough to meet AI-era electricity demand.