2 Green Flags and 2 Red Flags for Nuclear Stocks After This Year's Sell-Off

U.S. electricity demand is projected to rise 60% from 2025 to 2045, versus only 10% growth over the prior 20 years, supporting a long-term case for nuclear baseload generation amid AI and EV demand. NextEra's proposed acquisition of Dominion would expand its nuclear scale and capital-market access, while Constellation has signed deals tied to AI-driven power demand. However, the article flags unproven, cash-burning SMR developers such as Oklo and NuScale, high reactor construction costs, and elevated valuations including Constellation at 25x earnings versus roughly 19x for the average utility and Cameco at 157x.

Analysis

The investable scarcity is not nuclear technology broadly, but dispatchable, already-permitted generation located inside constrained power markets. CEG retains the clearest earnings torque to hyperscaler contracting and power-price repricing, but its premium makes it vulnerable to any evidence that AI load forecasts are slipping or that customers secure cheaper gas-backed capacity. NEE/D offers a lower-beta route to the same load-growth theme: scale lowers financing cost and should improve the ability to bundle generation, transmission and retail solutions, although regulated-return mechanics mean value realization will be measured in years rather than quarters.

The more important second-order constraint is grid interconnection and transmission, not reactor availability. Data-center developers unable to obtain firm power will pay for existing capacity, transmission upgrades, gas peakers and storage; this broadens the beneficiary set toward PWR and EME rather than concentrating upside in speculative SMR developers. It also caps the near-term addressable market for OKLO and SMR: even a favorable licensing outcome does not solve first-of-a-kind construction, fuel procurement, customer creditworthiness and project-finance execution simultaneously.

Near-term nuclear sentiment remains reflexive, so a 1-3 month correction in SMR-linked equities can continue without changing long-duration load fundamentals. The key 6-18 month catalyst is signed, creditworthy power-purchase agreements with disclosed pricing, delivery dates and collateral—not government awards or nonbinding MOUs. Falsification for the incumbent thesis would be falling forward power prices in key data-center regions, material hyperscaler capex cuts, or adverse state/federal decisions on plant life extensions; for SMRs, another equity raise before a binding financed project would confirm dilution dominates optionality.

Key Decisions for Investors

• Prefer NEE over CEG on a 6-18 month horizon: accumulate NEE on utility-sector weakness and hedge with a modest short CEG only if the valuation spread remains elevated. NEE has lower single-asset/power-price exposure and greater financing flexibility; exit the relative trade if CEG secures additional long-dated PPAs at economics materially above current fleet assumptions.
• Avoid directional long OKLO and SMR until a binding customer contract includes project financing, delivery schedule and fuel path. Treat disclosed cash runway and quarterly operating cash burn as the gating data; a financing announcement without these terms is a sellable sentiment spike, not fundamental validation.
• Express the load-growth bottleneck through a 6-12 month long PWR or EME basket rather than additional nuclear beta. Transmission and electrical-infrastructure orders monetize regardless of whether incremental firm supply is nuclear, gas or storage; reassess if utility capex guidance or large-load interconnection queues weaken.
• For existing CEG exposure, use 3-6 month downside protection around earnings or major PPA announcements rather than adding spot. The asymmetry is unfavorable at a premium multiple: a positive contract is partly expected, while weaker realized power pricing or lower load forecasts can compress both earnings expectations and the multiple.