2 Nuclear Stocks Worth Buying as AI and Geopolitics Drive Demand for Power

AI data center electricity demand is projected to more than double by 2030 to 945 TWh, strengthening the investment case for nuclear power suppliers. BWX Technologies is highlighted as a profitable, backlog-rich nuclear components supplier with about $8.6 billion in backlog and new U.S. Navy contracts, while Oklo offers higher-risk upside through its unproven microreactor and fuel-recycling design. The piece is largely thematic and stock-specific commentary, with a constructive but speculative tone.

Analysis

The real trade is not “nuclear for AI” in the abstract; it is the widening gap between companies that monetize the permitting/build cycle now and those that monetize first kilowatt-hour later. BWXT sits in the sweet spot because its revenue is effectively backed by federal capex and submarine/defense demand, giving it a cash-flow bridge while the AI power thesis remains a call option. That makes it materially lower risk than pure-play advanced reactor developers, and any incremental SMR enthusiasm should compress its discount to industrial defense peers.

OKLO is the higher-beta expression, but the market is likely overestimating how quickly optionality converts to distributable earnings. The key second-order effect is financing: every delay in NRC commercialization increases dilution risk and pushes commercialization further into a higher-rate, tighter-capital environment. If the pilot program lands well, the stock can re-rate sharply on credibility alone; if it slips, the downside is less about technology and more about time decay and capital structure pressure.

GEV is an underappreciated beneficiary because it can capture the “picks-and-shovels for everyone” layer across conventional generation, grid equipment, and SMR supply chains. That broad exposure makes it a cleaner way to own rising power demand without relying on a single reactor design winning regulatory approval. BE is more of a hedge against the same problem: if nuclear permitting drags, distributed gas/fuel-cell backup wins share as a bridge solution for data centers that cannot wait years for grid interconnects.

Consensus is probably underweighting the fact that the near-term winners are the firms that reduce customer wait time, not the firms with the most elegant long-duration power thesis. The market may also be too linear in assuming AI load growth automatically translates into nuclear demand; in practice, interconnect constraints, water rights, and local permitting can divert spend into temporary generation, microgrids, and behind-the-meter systems first. That creates a multi-year sequencing trade, not a binary winner-take-all outcome.

Key Decisions for Investors

• Long BWXT on a 6-12 month horizon as the lower-volatility beneficiary of nuclear reindustrialization; risk/reward is favorable because defense cash flow caps downside while AI-SMR sentiment can still re-rate the multiple.
• Buy GEV on pullbacks as the best diversified way to express the power-capacity buildout theme; expect upside over 3-9 months from equipment, services, and SMR adjacency with less binary regulatory risk than pure plays.
• Treat OKLO as a tactical event-driven long only into clear regulatory/catalyst milestones; size smaller than BWXT/GEV because 12-24 month dilution and delay risk can overwhelm fundamental optionality.
• Consider a pair trade long BWXT / short OKLO to isolate commercialization risk vs. proven monetization; this should work if the market keeps rewarding backlog and government-backed cash flows over speculative reactor timelines.
• Use BE as a hedge or smaller relative-value long versus OKLO if you want exposure to AI power demand without waiting for NRC approval; if nuclear timelines slip, BE’s near-term deployment path should outperform.