NuScale (SMR): the only approved SMR design, down 84%, and the customer it still lacks
NuScale holds the only NRC-approved small modular reactor design in the US and $1.9B of cash, yet booked $75,000 of revenue last quarter. What it sells, why the stock fell, and what would change the read.
The standard $SMR story is that NuScale won the race: it holds the only small modular reactor design the US regulator has approved, and AI data centers need exactly the firm, carbon-free power it promises. The first half is true. The second half is where the stock went wrong.
A regulatory approval is a gate, not a customer. NuScale booked $75,000 of revenue in the second quarter of 2026, and the share price closed at $8.42 on 2026-09-25, 84% under its $53.43 closing high. This piece walks through what NuScale actually sells, why an approved design has not turned into orders, what the balance sheet buys, and the one signed contract that would change the read.
Why it matters now
On 2026-09-11 UBS cut NuScale to Sell with a $6 target, arguing faster rivals are catching up; the stock fell 15.7% that day and law-firm investigations followed. The desk published a full teardown the same month: NuScale Stock (SMR): The Only Approved SMR, Down 84%, which is embedded on /stocks/smr with the live levels.
The TL;DR. NuScale designs a 77 MWe light-water reactor module and sells modules and engineering into projects led by its exclusive partner, ENTRA1 Energy. It owns the only NRC-approved SMR design in the US and about $1.9B of cash, but has no firm order: the 6 GW TVA program is still at the power-purchase-agreement stage, and the partner model already cost a $507.4M milestone payment.
What does NuScale do?
NuScale Power develops the NuScale Power Module, a small light-water reactor built in a factory rather than on site. Its current design, the US460, produces 77 MWe per module, and a standard plant stacks six modules for 462 MWe. Light water is the same basic technology as the large reactors already running in the US, which is why the regulator could approve it: the novelty is the size and the factory build, not the physics.
The pitch is repeatability. Large nuclear plants have become bespoke megaprojects that run years late and billions over budget. A module that rolls off a production line, ships, and bolts into a plant is meant to turn nuclear into a product with a learning curve.
What NuScale does not do is build, own or operate plants. That distinction decides almost everything about how it makes money.
How NuScale makes money
NuScale's future revenue is module sales and engineering services into projects that someone else develops. Since 2025 that someone is ENTRA1 Energy, its exclusive commercialization partner: ENTRA1 finds the customers, structures the projects and owns the plants, and NuScale is paid when a project orders modules.
Q2 2026 revenue: $75,000. The Romanian RoPower project (six modules at Doicesti) finished its front-end engineering phase in late 2025, and no firm module order has replaced it. On revenue, NuScale is a pre-commercial company with a $3.5B market value.
The model keeps NuScale capital-light on paper, but it is not free. In the fourth quarter of 2025 NuScale paid ENTRA1 a $507.4M milestone payment, which drove an operating loss of roughly $690M for the quarter. Under the agreement, NuScale owes ENTRA1 fees without guaranteed revenue, and investor class actions now question how the arrangement was presented.
That is the structural difference with $OKLO, the other listed reactor developer: Oklo plans to build, own and operate plants and sell the electricity, while NuScale sells the hardware and depends on a partner to land the buyer. The full comparison is in Oklo (OKLO): what it does, how it plans to make money, and the licensing risk.
The one gate NuScale has passed
In May 2025 the NRC granted Standard Design Approval to the US460. It remains the only small modular reactor design with that approval in the United States. An older 50 MWe version had been approved in 2020 and certified in 2023, but the uprated 77 MWe design is the one that matters commercially.
For competitors, licensing is still ahead: X-energy, TerraPower, GE-Hitachi and Oklo all face multi-year reviews before they can offer a licensed product. NuScale is past that gate. The question the market is now asking is whether being first through the regulator is worth anything if nobody has signed to buy the product.
The 6 GW pathway that is not signed
The volume story is the Tennessee Valley Authority. In September 2025 ENTRA1 and TVA announced a program for up to 6 GW of NuScale plants, roughly thirteen full six-module plants. For a company with no revenue, that is the whole bull case in one line.
The status is what to read carefully. At the Q2 2026 update the program was still at the power-purchase-agreement discussion stage. An announced program is an intention; a signed PPA is a customer. Until the second exists, the 6 GW is optionality, not backlog.
The numbers
Data as of 2026-09-25 for prices and ratios, and 2026-06-30 for the balance sheet. Check /stocks/smr for the refreshed figures.
| Metric | Value |
|---|---|
| Last close | $8.42 (2026-09-25) |
| From the closing high | -84% (high $53.43) |
| 1 year | -77.7% |
| Market cap | about $3.5B |
| Cash and investments | $1.9B (2026-06-30) |
| Net equity raised, H1 2026 | $984.5M |
| Operating cash outflow, H1 2026 | $372.9M |
| Q2 2026 revenue | $75,000 |
| Q2 2026 operating loss | $64M |
| Street | Hold, 15 analysts, mean target about $12 |
More than half of the market value is cash, which is the strongest line in the bull case. Where that cash came from is the strongest line in the bear case: close to a billion dollars of new equity in six months, paid for by existing holders through dilution.
Where NuScale sits in the nuclear trade
QA files NuScale in the Nuclear / SMR bubble. On 252 trading days of correlation with market beta stripped out, it moves most with its own crowd: Oklo at 0.82, $NNE (Nano Nuclear) at 0.78 and $LEU (Centrus) at 0.75. The fuel side follows at $UEC 0.69.
The telling detail is the next pair: D-Wave ($QBTS) and Rigetti ($RGTI), both at 0.70. The market prices NuScale next to quantum-computing stocks, as a story on future adoption, not next to utilities that sell power today. That is why it falls hard in a sentiment drawdown, whatever its own news. For the fuel-chain view of the same trade, see Cameco (CCJ) explained.
The bull case
- The only approved design. Competitors still face years of NRC licensing; NuScale's regulatory risk on the design is already retired.
- A named pathway to scale. The ENTRA1 and TVA program targets up to 6 GW, about thirteen plants, with the largest public power provider in the US.
- Cash to wait. $1.9B of cash and investments covers more than two years at the H1 2026 burn rate.
- Proven technology family. A light-water design avoids the fuel and materials questions that faster, more exotic reactor designs still have to answer.
- Priced near the lows. At $8.42 the stock sits 11% above its $7.59 closing low, so the market already prices a long wait.
The bear case
- No firm order. Revenue was $75,000 last quarter, RoPower finished its engineering phase without a module order, and analysts have floated a slip of the Romanian project to 2034.
- The partner model costs cash first. The $507.4M milestone payment to ENTRA1 came before any plant earned NuScale a dollar, and the arrangement is now the subject of investor litigation.
- Burn and dilution. $372.9M of operating cash left in H1 2026. UBS estimates about $700M of negative free cash flow over 2026 to 2028; other estimates run nearer $1B.
- The first-mover edge is shrinking. UBS argues Oklo, X-energy, TerraPower and GE-Hitachi are catching up, and Fluor, the former majority owner, sold its entire stake by April 2026.
- Story-stock trading. A 0.70 correlation with quantum names means NuScale reprices with speculative sentiment, not with its own milestones.
How to access NuScale
$SMR lists on the NYSE, so any broker with US-market access can hold it. Broad utility or clean-energy funds give almost no exposure to this specific thesis; the ETF holdings table on /stocks/smr lists the funds that do carry it, with their weights. To hold US-listed names from outside the US, see /stack/ibkr for the broker QA uses for direct exchange access.
The desk's full read, with the level map, peer table and verdict in one document, is the NuScale Teardown Dossier, included in Pro.
What to watch
- A signed TVA power-purchase agreement. The single input that converts the 6 GW program from intention into a customer. Nothing else on the list matters as much.
- A firm module order, anywhere. RoPower or any other project moving from engineering to a purchase order would be the first real revenue line.
- The Q3 2026 report. Cash, burn and any new ENTRA1 payments; the balance sheet is the clock on how long NuScale can wait.
- The $7.59 closing low and the $9.07 50-day average. A close below $7.59 leaves no support on the chart; a close back above the 50-day would be the first sign the tape stopped selling the story.
- Bubble correlation. If NuScale's link to the quantum names weakens while it holds with Oklo and Centrus, the market is starting to price it as a nuclear business rather than a sentiment trade.
Bubble shifts and rule-based alerts on $SMR and its nuclear peers are part of /pro.
Live data on this ticker: /stocks/smr - price, ETF holdings, bubble correlation, bot positions.
Bubble context: /bubbles/nuclear-smr - the cluster this name belongs to and how it's moving.
QuantAbundance is educational research. Nothing here is investment advice. See /disclosures.
The full dossier
Included in ProThe thesis, the numbers, the bear case and the level map for SMR, in one read. The deep version of what you just read.
See the dossierRelated bubbles
A mesa por trás deste artigo
Algumas destas são relações de afiliação. Ver as divulgações.
Related research
Go deeper
Get the daily digest.
One email a day · alerts + bubble shifts + new research. Free during beta.
Free when you join: the $SMR Teardown Dossier (NuScale Power), the full desk file we sell in the shop. One per new reader.
No spam. One email per day max. Pro adds Telegram trade alerts and higher AI-assistant limits.