Back to News
Market Impact: 0.42

Load Up on Nuclear Before the Data Center Energy Race Accelerates: These 3 ETFs Cover Reactors, Uranium, and Smart Grid

Artificial IntelligenceEnergy Markets & PricesCommodities & Raw MaterialsTechnology & InnovationInfrastructure & DefenseRenewable Energy TransitionCompany FundamentalsInvestor Sentiment & Positioning

The article argues that AI data-center power demand could reach up to 12% of U.S. electricity demand by 2028, making nuclear a key beneficiary across reactors, uranium fuel, and regulated utilities. It highlights three ETFs—NUKZ, URA, and NLR—with recent 1-year returns of 42%, 62%, and 37%, respectively, and notes URA’s 5-year and 10-year gains of 184% and 416%. The setup is constructive for the nuclear value chain, but returns will vary sharply by exposure: URA is the most volatile, NUKZ is the most concentrated, and NLR is the most defensive.

Analysis

The market is beginning to price nuclear less as a standalone commodity theme and more as a bottleneck solution to AI-driven load growth. The second-order winner is not just reactor owners but any balance-sheet-heavy firm that can sign long-duration power contracts and monetize grid scarcity; that favors CEG and PEG over more speculative developers because they can translate demand into contracted cash flows now, not after licensing cycles. BWXT also gains optionality as a “pick-and-shovel” supplier, which is typically where the cleanest risk-adjusted exposure sits once a theme moves from narrative to procurement.

The key asymmetry is between capital intensity and time-to-revenue. Hyperscalers need power in the next 12-36 months, while incremental nuclear supply often lives on a 5-10 year arc, so the near-term trade is really a scarcity premium on existing dispatchable assets and fuel, not a full discounting of future SMR capacity. That argues for CCJ as the highest-beta medium-term exposure if supply tightens, but also means the most crowded part of the trade is likely the reactor-builder narrative, where execution slippage can hit multiples before any megawatt is delivered.

The main contrarian risk is that the market may be overestimating how much of AI load growth gets solved by new nuclear versus demand-side management, gas peakers, grid upgrades, and behind-the-meter generation. If electricity price politics intensify, regulators could force more of the economics back to utilities, compressing upside in pure thematic vehicles while leaving regulated names relatively intact. The first reversal signal would be a slowdown in hyperscaler contracting or a broad reassessment of data-center capex, which would hit the SMR cohort fastest and the uranium miners only with a lag.