Not Oklo. Not NuScale. This Nuclear Stock Is Going to Be the Biggest Winner of the AI Power Boom.

Vistra is presented as a lower-risk way to gain exposure to AI-driven electricity demand versus pre-revenue small-reactor developers Oklo and NuScale. Vistra has 44 GW of capacity, including 6.4 GW of nuclear generation following its 2024 Energy Harbor acquisition; analysts forecast 2025-2028 revenue and EPS CAGRs of 13% and 75%, respectively. Despite risks from the 2025 Moss Landing battery-storage closures and PJM capacity-price caps through 2030, the stock is down more than 30% over 12 months and trades below 16x current-year earnings.

Analysis

The relevant valuation debate is not nuclear versus non-nuclear; it is whether dispatchable power in constrained load pockets can monetize AI demand before new supply arrives. VST has near-term cash flows and a mixed fleet, but its upside is concentrated in ERCOT and PJM market design, gas prices, hedge rolls, and capacity-price outcomes—not a generic data-center narrative. The PJM collar creates a multi-year ceiling on one source of scarcity rent, so consensus EPS growth should be stress-tested against realized auction clearing prices and plant availability rather than headline load forecasts.

OKLO and SMR remain duration-heavy development equities: their equity value is principally a function of licensing, fuel supply, customer financing, and construction execution years before meaningful operating cash flow. Higher long-end yields, an NRC timeline slip, or a single cancelled offtake agreement can drive material multiple compression even if data-center power demand remains robust. A second-order beneficiary of continued nuclear enthusiasm is uranium exposure (CCJ, UEC) and established nuclear operators (CEG), where fuel-cycle tightness and contracted demand can be monetized on a materially shorter timetable.

Contrarian view: VST's drawdown may have discounted some operational and regulatory uncertainty, but calling it inexpensive on forward P/E alone is incomplete because power earnings are inherently commodity- and weather-sensitive. The better setup is to own proven generation selectively while shorting pre-revenue reactor beta; that isolates the execution-duration premium likely to unwind if rates remain elevated. Over the next 1-3 months, quarterly guidance on hedged generation, Moss Landing remediation costs, and PJM/ERCOT forward curves are more important catalysts than incremental AI capex announcements; over 6-18 months, interconnection delays and gas-basis volatility determine whether power scarcity persists.

Key Decisions for Investors

• Initiate a 6-12 month pair: long VST / short a basket of OKLO and SMR, sized beta-neutral. The thesis is cash-generating dispatchable capacity versus permitting-and-financing duration risk; reassess if OKLO secures fully financed, binding commercial deployment or SMR announces a funded construction start with a creditworthy offtaker.
• Use a staged VST entry around earnings rather than chase AI-power headlines; add only if management reaffirms normalized free-cash-flow and hedge assumptions after remediation updates. Target 15-20% upside on earnings/FCF derisking versus 10-12% downside if PJM forwards weaken or outage/remediation costs rise materially.
• For cleaner nuclear exposure, prefer CEG over speculative SMR developers on a 6-18 month horizon, subject to valuation discipline. Monitor realized power prices, nuclear uprate economics, and new long-duration data-center contracts; a broad power-price decline or unfavorable contract repricing would invalidate the relative thesis.
• Set alerts for PJM capacity-auction rule changes, ERCOT forward-power deterioration, and 10-year Treasury yields moving sharply higher. These are likely to move VST and reactor developers more than general AI sentiment, with higher rates particularly negative for OKLO and SMR.