NuScale Power booked $75,000 of revenue last quarter. Not $75 million – seventy-five thousand dollars, down from $8.1 million a year earlier, per the company’s Q2 2026 10-Q. Two days ago, on 11 August, it registered to sell up to $750 million of new stock through an at-the-market programme. That is ten thousand times its quarterly revenue, and it is roughly 18% of the entire company. The shares closed at $9.59 on 12 August, down 3.03%, per stockanalysis.com. Our scenarios put a bull case at $18.00 (+88%), a base case at $10.00 (+4%) and a bear case at $4.50 (−53%).
The at-the-market filing is the part almost nobody has processed yet, and it changes how the next twelve months should be read. NuScale already raised $984.5 million of equity in the first half of 2026. The new sales agreement filed on 11 August appoints UBS, B. Riley, Canaccord Genuity, Craig-Hallum, Texas Capital Securities and Tuohy Brothers to sell shares at the company’s sole discretion, at up to 2% commission. At $9.59, $750 million is about 78 million new shares against the 429.7 million outstanding. An ATM sells into strength by design – which means every rally from here has a seller standing behind it, and that is a structurally different situation from the one the stock had a week ago.
Key facts
- $9.59 – SMR close, 12 August 2026, −3.03%; −41.2% year to date and −83.3% below its 52-week high – stockanalysis.com
- $75,000 – Q2 2026 revenue, down from $8.1m a year earlier – NuScale Q2 10-Q
- $47.5m – Q2 net loss attributable to Class A holders, $(0.13) per share; $91.6m across the first half – NuScale Q2 10-Q
- $750m – at-the-market equity programme registered 11 August 2026, about 18% of market capitalisation – NuScale 8-K, 11 Aug 2026
- $984.5m – equity already raised in the first half of 2026 – company disclosure
- $1.9bn – cash, equivalents and investments at 30 June, with no debt outstanding
- $372.9m – net cash used in operating activities in H1 2026, implying roughly 2.5 years of runway at that pace
- 429.7m – shares outstanding (410.4m Class A plus 19.3m Class B) – NuScale Q2 10-Q
What the quarter actually said
The revenue collapse has a benign explanation and a real one. The benign version: the $8.1 million booked a year ago came from Fluor’s FEED Phase 2 engineering work on the RoPower project in Romania, which completed in late 2025, leaving no comparable activity in the quarter. Revenue at NuScale is lumpy engineering recognition, not product sales, so a quarter near zero is mechanically possible without anything going wrong.
The real version is that this is what a pre-commercial company looks like eighteen years after founding. Net loss attributable to Class A shareholders was $47.5 million for the quarter and $91.6 million for the half. Accumulated deficit stands at $824.4 million. Operating cash burn was $372.9 million in six months, and management attributes it to technology readiness, design maturity, organisational growth and supply-chain preparation – all of which are investments in a product that has not yet been sold at commercial scale.
The genuine progress is real but preliminary: continued discussions with the Tennessee Valley Authority toward a possible power purchase agreement, ongoing work with Nuclearelectrica and RoPower on Romania’s Doicești project, and a contract awarded to Paragon to finalise the HIPS design, a supply-chain step for the NuScale Power Module. None of those is a signed order. A PPA under discussion is not a PPA.
The balance sheet is the whole story, in both directions
NuScale holds $1.9 billion in cash and investments with no debt. Against a market capitalisation of about $4.1 billion at $9.59, that is 46% of the equity value sitting in cash, or roughly $4.42 per share. Strip it out and the operating business carries an enterprise value near $2.2 billion.
That cash is why this is a $4.50 bear case rather than a $1 one. It is also why the $750 million ATM is so consequential. Two and a half years of runway is comfortable; the company did not need to file this now. Registering it anyway says management wants the option to fund toward a first commercial build without waiting for a marketed offering – or that it expects the burn to rise. Either reading points the same direction on share count.
Run the arithmetic. Draw the full $750 million at around current prices and the count goes from 429.7 million to roughly 508 million shares, an 18.2% increase. Cash per share falls even as total cash rises, because the numerator grows more slowly than the denominator once you net off a year of burn. This is the same mechanism that made BigBear.ai’s revenue per share fall 24% while revenue grew 13%, applied to a company with no revenue to divide at all.
The bull case: $18.00
The bull case requires one thing: a signed, financed, commercial order.
If the TVA discussions convert into an actual power purchase agreement, NuScale stops being a design company and becomes a company with a customer, a delivery schedule and a financeable backlog. That is a binary re-rating, not an incremental one. The market would move from valuing $1.9 billion of cash plus optionality to valuing a build programme, and the comparable set changes entirely.
At $18.00, the market capitalisation on the current share count is about $7.7 billion, implying roughly $5.8 billion of enterprise value for a first-of-a-kind SMR order book. That is demanding but not unprecedented for a company that has the only small modular reactor design with US Nuclear Regulatory Commission approval – the single genuine moat in the story, and one that took years and hundreds of millions to obtain. Add the AI power-demand backdrop, and a first order would arrive into the most receptive market for new nuclear capacity in forty years.
$18.00 is about 88% above spot and still 66% below the 52-week high. The bull case does not need a mania. It needs a signature.
The regulatory position deserves more weight than the market currently gives it. Design approval from the NRC is not a formality that competitors can replicate on a similar timeline – it is a multi-year, multi-hundred-million-dollar process, and it is the reason NuScale cannot simply be leapfrogged by a better-funded entrant. Where the company sits against the alternatives is worth setting out plainly:
| NuScale (SMR) | Oklo (OKLO) | Operating nuclear (CEG) | |
|---|---|---|---|
| Revenue today | $75,000 last quarter | $1.2m first revenue | Billions, from 21 reactors |
| Regulatory status | NRC-approved design | Earlier-stage licensing | Already operating |
| What you own | An option on a first order | An option on a first build | Cash flows and PPAs |
| Cash position | $1.9bn, no debt | Smaller, still raising | Utility balance sheet |
Read down the first column and the investment case is coherent: the best regulatory position in the category, the strongest balance sheet among the pre-revenue names, and no customer. Read across the row marked “revenue today” and the risk is equally clear. NuScale has spent its lead time accumulating approvals rather than orders, and approvals do not amortise – they expire into irrelevance if nobody builds.
The bear case: $4.50
The bear case is that the ATM gets used and the order does not arrive.
Assume the full $750 million is drawn over the next eighteen months and the company keeps burning roughly $750 million a year. Cash goes from $1.9 billion to something closer to $1.9 billion again – the raise funds the burn rather than adding to the balance sheet – while the share count rises to about 508 million. Net cash per share lands near $3.74. A market that has lost patience with pre-revenue nuclear prices the equity at a modest premium to net cash, which is roughly $4.50, about 53% below spot.
Two things make that scenario more plausible than it sounds. First, this stock has already fallen 83.3% from its 52-week high and 41.2% year to date, so the de-rating is not hypothetical – it is in progress. Second, the timeline risk is asymmetric: SMR construction schedules slip as a matter of industry routine, and each slip pushes first revenue further out while the burn continues at the same rate. The peer read-across is unflattering too – Oklo’s first revenue was $1.2 million against a $48.5 million loss, and that stock is down 42% year to date.
The base case: $10.00, and what you are actually buying
Our base case of $10.00 is roughly 4% above spot, and the honest description is that the market has this approximately right. You are paying about $4.1 billion for $1.9 billion of cash and a regulatory approval, with a $750 million issuance authorisation now hanging over it.
That is a coherent thing to own, provided the buyer understands what it is. SMR is not a power company and not an energy stock in any conventional sense – it generates no electricity and sells no product. It is a long-dated option on the first commercial small modular reactor order in the United States, funded by a cash pile that is being consumed at roughly $750 million a year and topped up by issuing stock. Sized as an option, it is defensible. Sized as an energy holding, it is mispriced by category error.
For the profitable end of the same AI-power theme, see our analysis of Bloom Energy’s bull and bear cases, where revenue grew 165% and the company is already generating GAAP net income. The contrast between the two is the clearest illustration available of what “AI power stock” can mean at opposite extremes.
What moves the number next
The TVA power purchase agreement. This is the binary. A signed PPA is the difference between the bull case and the base case, and there is no partial credit – discussions have been ongoing for some time without conversion.
The ATM drawdown rate. Watch the share count in the Q3 filing. If NuScale has sold meaningfully into this bounce, the market learns that management treats $9-10 as an acceptable issuance price, which caps the stock at that level far more effectively than any analyst target.
RoPower reaching a final investment decision. Romania’s Doicești project is the furthest advanced. An FID there would be the first real proof that a NuScale module gets built by somebody.
Any move in the burn rate. The $372.9 million consumed in six months is the clock everything else runs against. Management has framed the spending as readiness investment, which is defensible while a first order is plausibly close and indefensible if it is not. A burn that rises again in Q3 without a corresponding commercial milestone would tell investors the company is scaling for demand it has not yet secured, and that is the configuration in which the ATM stops being optional.
Our base expectation is that NuScale spends the next two quarters between $8 and $12, funding itself through the ATM while the TVA talks continue, with the stock reacting violently in both directions to headlines rather than results. This is a name where the news flow, not the financials, sets the price – and with $75,000 of quarterly revenue, that is simply the accurate description.
This analysis is for information only and is not investment advice. Scenario prices are FinanceFeeds estimates derived from net cash, share count and order-book assumptions, and are not price predictions or recommendations. Do your own research.
Frequently asked questions
What is the SMR stock forecast for 2026?
Our scenarios put NuScale’s bull case at $18.00, base case at $10.00 and bear case at $4.50, against a spot price of $9.59 on 12 August 2026. The outcome hinges almost entirely on whether the Tennessee Valley Authority discussions convert into a signed power purchase agreement.
How much revenue does NuScale actually make?
Almost none. Q2 2026 revenue was $75,000, down from $8.1 million a year earlier, because the Fluor FEED Phase 2 engineering work on Romania’s RoPower project completed in late 2025 and no comparable activity replaced it. NuScale is a pre-commercial company; its revenue reflects engineering milestones, not product sales.
What is NuScale’s $750 million ATM programme?
On 11 August 2026 NuScale registered an at-the-market offering allowing it to sell up to $750 million of Class A stock at its discretion through six sales agents at up to 2% commission. At $9.59 that is roughly 78 million new shares, about 18% of the 429.7 million outstanding. ATM programmes typically sell into price strength.
Is NuScale running out of money?
No, not in the near term. It held $1.9 billion in cash and investments at 30 June with no debt. First-half operating cash burn was $372.9 million, implying roughly 2.5 years of runway before the new ATM. The concern is dilution, not insolvency.
Why has SMR stock fallen so far?
The stock is down 41.2% year to date and 83.3% from its 52-week high, having traded above $53 in late 2025. The decline reflects the gap between nuclear enthusiasm and commercial reality: no signed orders, revenue near zero, continuing losses and repeated equity issuance while first deployment remains years away.
Is NuScale a good way to invest in AI power demand?
It is the longest-dated way. NuScale generates no electricity today and its reactors are unlikely to be operating before the late 2020s, so it cannot serve current AI power demand at all. Investors seeking exposure to that demand now are looking at companies already delivering power, whereas NuScale is an option on the next decade.