The Last U.S. Nuclear Boom Applied to Build 26 Reactors and Finished 2. NuScale Carries a $4 Billion Market Value on $10.7 Million of Sales.

NuScale reported Q2 revenue of just $75,000 (down from $8.1M a year earlier) and a $47.5M Class A shareholder net loss, while H1 2026 revenue totaled $640,000. The company also announced a new $750M at-the-market stock sale and has $1.9B cash/investments, but trailing sales of $10.7M versus a ~$4.2B valuation suggests the market is pricing future reactor orders rather than current performance. The article highlights a long wait for signed, funded orders, contrasting the “nuclear renaissance” (26 proposed reactors, only Vogtle completed) with NuScale’s factory-built SMR pitch and ongoing talks/pending projects (TVA discussions; Romania conditions for RoPower).

Analysis

SMR is still being valued as if it has already solved the hardest part of nuclear: converting policy interest into bankable, timetable-specific cash flows. The real bottleneck is not technical certification; it is who underwrites construction, who absorbs delay risk, and whether an off-taker will sign a contract that survives diligence and financing. In that framework, the current equity value looks less like a balance-sheet-backed industrial and more like a venture claim on future order flow, which makes it highly sensitive to any slippage in deal conversion or to a single financing round that resets per-share economics.

Second-order, the supply chain story cuts both ways. Factory fabrication and partner counts can reduce site risk, but they also push execution complexity upstream into module manufacturing capacity, QA, and schedule coordination; any miss there still creates multi-quarter delays, just earlier in the process. If data-center buyers are the real new demand pool, they are likely to demand completion guarantees and fixed-price structure, which would compress economics for the developer and shift risk onto suppliers, EPCs, and financiers rather than disappear it.

The near-term catalyst path is binary: a signed, funded power agreement or a capital raise that reinforces the dilution overhang. Without hard contracts, the market can quickly re-rate this from "scarcity option" to "cash-burning pre-revenue story"; with a contract, the debate shifts to whether NuScale can deliver multiple modules on time and on budget, which is a multi-year proof point. The contrarian miss is that enthusiasm for AI power demand may be real, but scarcity does not automatically equal monetizable moat when customers can wait for lower-cost alternatives or negotiate down economics.

The main falsifier is a definitive PPA plus project financing that is non-dilutive and includes credible timing milestones; absent that, the thesis remains that headline optionality is outrunning execution probability.

Key Decisions for Investors

• Short SMR on strength into any order-related headline; use a tight stop above the prior gap-up level because the stock will likely trade as a financing/catalyst binary.
• If the name rips on TVA or Romania progress, sell 30-50% of the move and wait for the financing terms; the highest-probability upside is already embedded in a signed, funded contract, not a memorandum of understanding.
• Consider a bearish put spread on SMR with a 1-3 month horizon to express the risk that ATM usage or a secondary offering becomes the real catalyst instead of backlog conversion; structure it to pay off if the market starts pricing another dilution round.
• Watch for a definitive PPA announcement and project finance terms as the thesis breaker; if those arrive, reassess immediately because the market would then move from narrative risk to execution risk.
• If looking for a cleaner relative-value expression, pair short SMR against a basket of profitable power/utility exposure (XLU or a utility with visible regulated earnings), since SMR’s multiple is most vulnerable to a reset in "story premium" while utilities benefit from actual contracted cash flow.