Uranium ETFs Just Crashed 30 Percent While AI Power Demand Keeps Breaking Records

David Beren

Tue, August 18, 2026 at 7:41 PM GMT+3 6 min read

Quick Read

URA and URNM have fallen roughly 30% from their highs even as AI-driven data centers race toward 12% of U.S. electricity demand by 2028.
NLR's nuclear utility holdings, including Constellation Energy and PSEG, capture AI power purchase agreements without depending on uranium spot prices at all.
URNM's near-50% concentration in Cameco, NexGen, and physical uranium makes it the highest-beta ETF for investors betting on a sharp spot price recovery.
Don't wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.

Uranium equities have given back roughly a third of their value since last year's highs, even as hyperscaler power contracts, small modular reactor announcements, and utility load forecasts continue to point in one direction. The three funds most investors use to play the theme, Global X Uranium ETF (NYSEARCA:URA), Sprott Uranium Miners ETF (NYSEARCA:URNM), and VanEck Uranium and Nuclear ETF (NYSEARCA:NLR), have all felt the pressure, but the pain has not been evenly distributed.

24/7 Wall Street

URNM's 52-week range runs from $44 to $65, a peak-to-trough drawdown that lines up with the roughly 30% correction narrative. URA sits at $45 after a 9% bounce over the past week, and NLR trades near $117. The three funds diverge in construction, and the gap between them is where the investment case lives.

Why the Selloff Collided With Record Power Demand

The uranium spot price weakened this year as term contracting slowed and a handful of production restarts came online. That happened while the underlying electricity story continued to accelerate. The Department of Energy projects that data centers will account for up to 12% of U.S. electrical demand by 2028, driven by AI. A single hyperscale facility can pull over a gigawatt of power, equivalent to powering approximately 750,000 homes.

Don't wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.

Nuclear generation is the one dispatchable, carbon-free source utilities can point to when a hyperscaler asks for round-the-clock baseload. That structural pull has not disappeared. It is why the equity drawdown looks disconnected from the operating environment, and why the three ETFs below express the theme in very different ways.

URA: The Default, With a Broader Net

Global X Uranium is the largest and most liquid uranium ETF, and for most investors, it is the reflex choice. Its portfolio blends uranium miners with companies further along the fuel cycle and the reactor technology curve. Recent disclosures show Cameco Corp at 22.11%, NexGen Energy at 6.10%, and Uranium Energy Corp at 4.81%, with smaller weights in reactor developers such as NuScale Power at 3.12% and X-Energy at 3.71%. That mix is what separates URA from a pure miner basket.

The fund carries an expense ratio of 0.69%, the lowest of the three. Over five years, URA has returned 178%, and it is up 17% over the past year, even after the correction. A $ 2.08-per-share distribution in January 2026, well above the prior January payout, reflected strong underlying cash generation despite equity weakness.

The tradeoff with URA is dilution. Adding reactor developers and fuel cycle names softens the fund's beta to the uranium price. That is useful in a downdraft, but it also mutes the upside when miners rally hard on a spot price move.

URNM: The High-Beta Pure Play

Sprott Uranium Miners is the concentrated version of the trade. The top three positions, Cameco at 20.69%, Sprott Physical Uranium Trust at 13.60%, and NexGen at 12.65%, account for nearly half the fund. Physical uranium exposure through the Sprott Trust means URNM tracks miners and holds the metal itself.

That construction is why URNM's drawdown looked steeper. It is down about 10% from early January and slightly negative year-to-date, while URA is roughly flat for the year. The other side of that concentration is leverage on the way up. Over the past week, URNM has climbed about 8%, and it is still up about 20% over the past 12 months.

The expense ratio of 0.75% is slightly higher than URA's. The relevant question is whether an investor wants the highest beta to the uranium spot price available in ETF form. URNM is that fund. If the thesis is a supply-driven price recovery in the metal, this is the sharpest tool. If the thesis is broader, it is the wrong one.

NLR: The Overlooked Utility Angle

The fund that most casual screens skip is NLR, and it is arguably the one that most directly captures the AI power demand story. VanEck's approach blends uranium miners with utilities that actually operate nuclear plants. The top three holdings are Constellation Energy at 9.42%, Cameco at 7.87%, and Public Service Enterprise Group at 7.86%, with the top ten representing 62.30% of assets.

Utilities like Constellation and PSEG earn revenue by selling electricity to end customers. When a hyperscaler signs a power purchase agreement with an existing nuclear operator, the operator benefits regardless of where uranium trades on the spot market that week. That decoupling is why NLR's year-to-date decline of about 6% is milder in character than the miner drawdowns, and why its one-year return of about 2% is more subdued in both directions.

More income is also what NLR pays. The fund offers a dividend yield of 2.7%, well above those of URA or URNM, reflecting its utility-heavy composition. Its expense ratio of 0.62% is the lowest of the three. The tradeoff is muted upside if uranium prices spike, since utilities do not re-rate the way developers do on a $10 move in spot U3O8.

Which Fund Fits Which Investor

The three funds map cleanly to three views of the theme. URA offers the standard, liquid uranium exposure for investors without a strong view on how the recovery unfolds. Its breadth of holdings and lower cost position it as the broader-market option.

A view that the recent sell-off overshot and that uranium spot prices are set for a sharp move aligns with URNM. The concentration in Cameco, NexGen, and physical uranium is the whole point. It will hurt more in the other leg down and pay more in a rally.

A thesis centered on AI power demand rather than uranium prices is a fit for NLR. Constellation and PSEG do not need spot uranium to rise to benefit from data center contracts and grid capacity constraints. The yield is a bonus. It is the least obvious pick on this list and, for that specific view, the most defensible one.