NuScale Power Stock Broke Out in August. Is It a Buy?

NuScale Power rose 10.1% in August after announcing nuclear-focused AI tools that could reduce engineering information-search time by up to 80%, but Q2 revenue collapsed 99% year over year to $75,000 following completion of Romanian design work. The company ended Q2 with $1.9 billion in cash and investments, up $900 million sequentially, and filed to sell up to $750 million of additional shares, creating further dilution risk. Although its NRC-approved SMR design could address AI data-center demand for carbon-free baseload power, its first reactor remains years from operation and the investment case hinges on converting non-binding agreements into signed PPAs and factory orders.

Analysis

SMR remains a duration-sensitive development asset rather than an AI-power beneficiary with measurable earnings leverage. The relevant valuation variable is not document-search productivity but the probability-weighted conversion of prospective customers into bankable, creditworthy PPAs; absent that, incremental engineering efficiency does little to offset corporate overhead, development spend, and future equity issuance. The financing overhang should cap sustained rallies because each price increase improves the issuer’s incentive to fund runway through stock, creating reflexive supply.

Near term (days to 1 month), the stock can squeeze on data-center power headlines, DOE/NRC milestones, or a customer framework announcement, but these are weak substitutes for contracted economics. Over 1-3 months, monitor whether ENTRA1 identifies project sites, offtakers, interconnection rights, construction financing, and an EPC/risk-allocation structure; any one missing link materially reduces the value of an announced PPA. Over 6-18 months, nuclear’s scarcity premium likely accrues more reliably to incumbents with operating fleets, fuel access, and regulated rate-base pathways—CEG and VST in U.S. power, and CCJ/LEU in fuel-cycle bottlenecks—than to a pre-revenue reactor developer.

The contrarian bullish case is that large-load customers may pay a premium for firm, carbon-free capacity as gas turbines, transmission equipment, and interconnection queues become binding constraints. But SMR is unlikely to be the marginal solution to near-term AI load growth: hyperscalers need capacity this decade, while first-of-a-kind construction, licensing amendments, supply-chain qualification, and project-finance terms can extend beyond promotional timelines. A signed PPA without disclosed price, escalation, collateral, and completion-risk allocation should not be treated as de-risking.

Key Decisions for Investors

• Maintain no core long in SMR pending a disclosed, binding PPA with customer credit support, project capacity, delivery date, and financing structure. A framework agreement or AI-tool update is an alert, not a buy trigger.
• On a headline-driven SMR rally, consider a 1-3 month tactical short or long put spread only if borrow is available and implied volatility is not prohibitive; target a retracement toward the pre-headline range, with stop discipline above the announcement-day high. Primary risk is a genuinely bankable hyperscaler-backed contract.
• Express the AI-power theme through long CEG and/or VST over SMR for a 6-18 month horizon: existing generation, contracted/repricing opportunities, and operating cash flow provide more direct exposure to firm-power scarcity. Size for regulatory and power-price volatility.
• For a higher-beta supply-chain expression, watch LEU and CCJ rather than initiating solely on this article. Enter only after confirmation of durable reactor-build demand or fuel-contract tightening; thesis is falsified by uranium enrichment/fabrication supply normalization and delayed reactor orders.
• Set a catalyst checklist for SMR’s next results: cash burn versus guidance, share-count growth, booked backlog conversion, and quantified customer commitments. A material burn acceleration or another large at-the-market issuance would reinforce the dilution-led downside thesis.