No US utility has ordered a newly designed commercial reactor and seen it come online since the 1970s. On April 24, 2026, the first company trying to break that drought rang the Nasdaq bell. X-energy priced its IPO at $23 a share — $4 above its marketed range — and raised $1.02 billion, the largest US nuclear public offering in years. Shares jumped 26% on the first day.
No US utility has ordered a newly designed commercial reactor and seen it come online since the 1970s. On April 24, 2026, the first company trying to break that drought rang the Nasdaq bell. X-energy priced its IPO at $23 a share — $4 above its marketed range — and raised $1.02 billion, the largest US nuclear public offering in years. Shares jumped 26% on the first day.
X-energy has not yet poured concrete on a single reactor. But Amazon has committed to buy 5 gigawatts of its output by 2039, Dow Chemical wants one beside a Texas petrochemical plant, and eleven more gigawatts sit in the order book. The IPO tests a simple question: will public markets fund the decade-long gap between a reactor deal and delivered electrons, so AI data centers can finally get the firm carbon-free power they have been promised?
Why it matters
If small modular reactors deliver on time, AI data centers get 24/7 carbon-free power; if they slip, that demand falls onto natural gas.
Key Indicators
$1.02B
X-energy IPO proceeds
Gross proceeds from the April 2026 Nasdaq debut, the largest US nuclear listing in years.
5 GW
Amazon's X-energy commitment
Gigawatts of Xe-100 capacity Amazon has agreed to buy by 2039 to power data centers.
11 GW
X-energy order pipeline
Gigawatts of reactor orders across Amazon, Dow, Centrica and others — none yet built.
80 MW
Per Xe-100 reactor
Each unit generates 80 megawatts; plants can stack up to twelve units for 960 MW.
2030s
First commercial operation
X-energy's first Xe-100 is targeted to start generating power in the early 2030s.
Shares debut on the Nasdaq Global Select Market and jump roughly 26% on the first day, delivering paper gains to Amazon, Ken Griffin, and Kam Ghaffarian.
X-energy upsizes and prices IPO at $23
Financial
X-energy prices 44.25 million shares at $23 — $4 above the top of its range — raising $1.02 billion, the largest US nuclear IPO in years.
X-energy files S-1 for up to $800M IPO
Filing
The company formally files to go public on Nasdaq, initially targeting $16–$19 per share and about $800 million in proceeds.
Meta signs 16-reactor Oklo deal for Ohio AI campus
Deal
Meta partners with Oklo to build a 1.2 gigawatt campus in Pike County, Ohio using 16 Aurora Powerhouse microreactors, with prepayments to secure HALEU fuel.
X-energy closes oversubscribed $700M Series D
Investment
Private-round fundraising aimed at reactor licensing and scaling the TRISO-X fuel plant in Oak Ridge, Tennessee.
Amazon unveils 'Cascade' Xe-100 project
Deal
Amazon, X-energy and Energy Northwest name their Washington project Cascade — a 320 MW four-unit build targeting construction by 2030 and first power in the early 2030s.
AWS signs 17-year PPA with Talen for Susquehanna
Deal
Amazon agrees to buy 1.92 gigawatts from Talen Energy's existing Susquehanna nuclear plant, committing $20 billion in Pennsylvania data-center investment.
NRC caps advanced reactor review at 18 months
Regulatory
Under guidelines prompted by the 2024 ADVANCE Act, the NRC commits to completing qualifying advanced reactor reviews within 18 months, down from a historical 5 to 7 years.
X-energy and Dow file NRC construction permit
Regulatory
The companies submit a construction permit application for four Xe-100 reactors at Seadrift — the first NRC application to build an advanced reactor at an industrial site.
Amazon leads $500M Series C-1 into X-energy
Investment
Amazon anchors a $500 million round with Ken Griffin, Ares, NGP, and the University of Michigan. Amazon separately commits to buy up to 5 gigawatts of X-energy capacity by 2039.
Microsoft signs 20-year PPA to restart Three Mile Island
Deal
Constellation Energy agrees to restart the retired Three Mile Island Unit 1 to sell all its output to Microsoft for AI data centers — reopening a reactor is easier than building one.
Amazon buys Talen's nuclear-adjacent Cumulus data center
Deal
AWS purchases the Cumulus data center campus attached to Talen's Susquehanna nuclear plant in Pennsylvania, signalling hyperscaler appetite for directly co-located nuclear power.
NuScale's UAMPS project collapses over cost
Setback
The Utah Associated Municipal Power Systems cancels the Carbon Free Power Project after cost estimates rise to $9.3 billion. The first US SMR deal fails before construction.
Dow and X-energy partner on first industrial-site SMR
Deal
Dow Inc. and X-energy agree to develop an Xe-100 reactor at Dow's Seadrift, Texas petrochemical complex, the first plan to site an advanced reactor at a commercial industrial facility.
NRC certifies NuScale's SMR, the first US approval
Regulatory
The Nuclear Regulatory Commission certifies NuScale Power's small modular reactor design, making it the only SMR cleared for US deployment.
DOE picks X-energy and TerraPower for Advanced Reactor Demonstration Program
Policy
The Department of Energy awards X-energy and TerraPower $80 million each in initial funding, with up to $1.2 billion apiece in cost-share support to demonstrate advanced reactors by 2027.
Scenarios
1
First Xe-100 comes online by 2031, order book doubles
Discussed by: X-energy S-1 filing; Bank of America nuclear analysts; Amazon sustainability disclosures
The NRC clears Dow's Seadrift permit within its new 18-month window, Oak Ridge fuel production ramps on schedule, and first power flows from a Texas Xe-100 before 2031. Commercial proof drives new hyperscaler, industrial, and allied-country orders, pushing X-energy's 11-gigawatt pipeline past 20 gigawatts and validating the SMR thesis for the rest of the sector.
2
Construction slips to late 2030s, hyperscalers default to gas
Discussed by: Skeptical nuclear analysts; past NuScale and Vogtle cost-overrun post-mortems
Licensing, HALEU fuel supply, or construction hits the same cost spiral that sank NuScale's Utah project and ballooned Vogtle 3 and 4 by $20 billion. Xe-100 first power slides to the late 2030s. To meet AI demand on time, Amazon, Meta, and peers sign more natural-gas and on-site generation deals; SMR equity values compress.
3
HALEU shortage bottlenecks the whole sector
Discussed by: US Department of Energy; Centrus Energy analysts; IAEA fuel-cycle reports
X-energy, Oklo, and TerraPower all rely on high-assay low-enriched uranium. The US only recently restarted domestic enrichment after decades of Russian dependence. If HALEU production scales slower than reactor demand, first cores arrive late even for well-licensed projects. X-energy's Oak Ridge TRISO-X plant becomes the most important node in the sector.
Discussed by: Industry M&A bankers; Reuters energy desk
Not every SMR startup survives. As public-market discipline arrives via IPO, weaker designs lose funding and get bought by GE-Hitachi, Westinghouse, or cash-rich hyperscalers. X-energy either consolidates peers or is itself absorbed by a large utility seeking vertical integration. Three-to-five designs survive globally by 2030.
Historical Context
NuScale UAMPS Carbon Free Power Project (2015–2023)
2015–November 2023
What Happened
Utah Associated Municipal Power Systems and NuScale Power spent eight years developing what was meant to be the first commercial US small modular reactor — a six-module plant at Idaho National Laboratory. After repeated redesigns, rising interest rates, and cost estimates that climbed from $3 billion to $9.3 billion, the project was cancelled in November 2023. Target power prices had risen to $89 per megawatt-hour.
Outcome
Short Term
NuScale's share price collapsed, workforce was cut sharply, and the company pivoted toward marketing its NRC-certified design to new customers.
Long Term
The cancellation reset expectations that small reactors would be cheap, and reinforced the view that SMR projects only work with deep-pocketed anchor offtakers rather than small public utilities.
Why It's Relevant Today
X-energy's economics only close because Amazon, Dow, and Energy Northwest are pre-committed to buying the power. The UAMPS failure is exactly the kind of outcome hyperscaler offtake is meant to prevent.
Vogtle Units 3 and 4 construction (2009–2024)
August 2009–April 2024
What Happened
Georgia Power began building two Westinghouse AP1000 reactors at Plant Vogtle in 2009, with original cost estimates near $14 billion and a first-power target of 2016. Unit 3 entered commercial operation in July 2023 and Unit 4 in April 2024. The final cost exceeded $35 billion — more than double the budget and roughly seven years late.
Outcome
Short Term
Georgia ratepayers absorbed billions in cost-recovery charges, and no other US utility ordered a large reactor.
Long Term
Vogtle became the reference case against which any new nuclear project is judged, and a key reason investors now prefer smaller, factory-built reactor designs.
Why It's Relevant Today
X-energy's pitch to the market — factory manufacturing, smaller modules, industrial or data-center offtake — is explicitly engineered to avoid a Vogtle-style cost spiral.
Westinghouse bankruptcy and V.C. Summer cancellation (2017)
March–July 2017
What Happened
Westinghouse Electric Company filed Chapter 11 bankruptcy in March 2017, crushed by roughly $9 billion in construction losses on Vogtle and on two AP1000 units at South Carolina's V.C. Summer site. In July 2017, SCANA and Santee Cooper abandoned V.C. Summer entirely after spending about $9 billion with nothing to show.
Outcome
Short Term
Toshiba, Westinghouse's parent, took multibillion-dollar writedowns; Westinghouse was later sold to Brookfield. Thousands of construction workers were laid off in South Carolina.
Long Term
The collapse functionally ended the large-reactor business model in the United States and opened the door for smaller reactor designs that could be financed and built in smaller chunks.
Why It's Relevant Today
The SMR thesis X-energy is monetising was born from this failure. Public markets backing a reactor developer again — a decade after Westinghouse broke — is the clearest sign that investor memory is shifting.
