The Smart Money Is Quietly Fading the SMR Trade
Four independent skeptics, one stretched valuation, and a retail crowd buying the opposite.
Small modular reactors have become the purest expression of the AI power trade. On June 18 the whole complex ripped again. On retail timelines, the SMR names sit at the top of nearly every generational wealth list. Strip out the hype, and the serious analysts are not nibbling at the edges. They are fading the trade outright, on four separate grounds. That gap between the crowd and the careful money is the story.
One day price change, June 18 session. The pre-revenue pure plays lead the move. Source: SwiftAlerts real-time quotes.
The valuation is the easy part
The nuclear names that carry actual earnings now trade at 45 to 97 times them. The pure plays cannot be valued on earnings at all, because there are barely any. NuScale changes hands at roughly 211 times sales with no profit. Oklo carries a negative forward multiple. These are options on a future, priced as though the future has already arrived.
Trailing price to sales. NuScale sits roughly sixteen times above the next nuclear name and earns no profit. Oklo has effectively no revenue to value against. Source: SwiftAlerts fundamentals, June 19 snapshot.
The picture holds across every lens. On enterprise value to EBITDA, the two names with positive cash flow sit at 51 and 83 times, while the pure plays are negative. There is no version of this table where the sector looks cheap.
The crowd owns these names as a generational wealth bet. The careful money is fading the premise underneath them.
The bear case, in four parts
What makes this setup unusual is not a single skeptic. It is that four independent arguments, from people who rarely agree on anything, all point the same way.
| # | Argument | The case |
|---|---|---|
| 01 | The shortage may not be real | The whole complex repriced to 50–70x earnings on the assumption of a structural US power shortage. One prominent short seller argues that premise is simply wrong, and the system is heading toward surplus, not scarcity. This is a bet against the thesis, not one ticker. |
| 02 | Free electrons win | Nuclear goes sideways while solar and wind go parabolic. The decisive variable is the marginal cost of solar reaching effectively zero over the next few years. Paying a premium for unproven reactors against near-free power is a hard case to defend. |
| 03 | Wrong layer, wrong decade | Legacy plants cannot ramp and SMRs remain stuck in regulation, so they miss the power bottleneck that actually binds in 2026–2028. The fast-deployers — gas, microturbines, on-site power — win that window instead. |
| 04 | Insanely expensive at scale | Outside China, the theoretical levelized cost of nuclear diverges sharply from what plants actually deliver. The new generation of SMRs is shaping up to be very expensive, well before a single commercial unit proves the cost curve. |
The largest pure play carries its own scars
Beyond the theme, the biggest listed SMR name has company-specific history worth isolating. It has faced direct regulatory scrutiny, including a finding that staff could not confirm a risk-informed process was in place. It was also the subject of an enforcement inquiry, followed by an awkward sequence in which the company first said it was unaware of the matter and then filed to clarify that it had, in fact, heard from the agency. None of that is dispositive, and much of it is dated, but it belongs in any honest underwriting of the ticker.
Illustrative sequencing. The solutions that deploy fast capture the 2026–2028 power crunch. The SMR names are early to a party that starts without them. Source: SwiftAlerts, synthesizing tracked research.
The steel man, because it is real
The bull case is not empty, and a fair note has to say so. The demand side is genuine. The original hyperscaler reactor deal — a major cloud provider agreeing to buy firm clean power from an SMR developer — predates the mania and reflects real corporate appetite. The strongest structural argument is the manufacturing one: the industry is trying to move from bespoke megaprojects to an assembly-line model, a learning curve that, if it works, drives cost down the way it did for every other modular product. And the regulatory picture is not frozen. One leading pure play has since cleared a preliminary federal safety milestone and signed a fuel letter of intent, which softens the worst of the licensing skepticism. The debate, properly framed, is about timing and economics, not about whether the electricity will be needed.
How to hold it
The honest synthesis is that the crowd and the careful money have rarely been this far apart on a single theme. Retail owns SMR names as a generational wealth bet. The skeptics question the premise, the economics, the timing, and the price per kilowatt-hour all at once, while the largest pure play carries regulatory and disclosure scars. That does not make the sector un-investable. It means the easy, narrative-driven phase of the move is the dangerous one, and that the better risk-reward sits in the parts of the power stack that solve the near-term bottleneck, or in waiting for the SMR names to reset toward something a fundamental model can defend.