Nuclear, Sentiment Correction Not Thesis Break
Uranium and nuclear equities have corrected 25 to 40 percent from their highs while the physical market has not moved. The divergence is a sentiment story, and a familiar one.
Uranium and nuclear equities have fallen 25 to 40 percent from their highs. The physical market has not followed. Spot uranium holds around $84 to 90 a pound, term prices sit above $90 with base escalated contracts signing in the mid $90s, and utility contracting continues. When the equities move that far and the commodity barely moves, the gap is telling you something about sentiment, not fundamentals.
What is driving the selloff
Three forces, and none of them is fundamental. The first is AI capex anxiety compressing the thematic premium. The sector rerated hard over 2024 and 2025 on the datacentre and nuclear narrative. As markets question the near term profitability of hyperscaler infrastructure spend, the premium embedded in names like Cameco, Oklo, Constellation and the small modular reactor complex is being handed back.
The second is a broad commodity and equity risk off. Sector specialists describe an air pocket largely unrelated to nuclear fundamentals. Announcements through the period were constructive, but the stocks were caught in a wider de risking across commodity sectors, and high beta thematic vehicles absorb the worst of that.
The third is crowded positioning unwinding. Retail flows chased the AI angle into the 2025 highs. The marginal buyer has since stepped away, and equities have corrected 25 to 40 percent across the ASX, TSX and NYSE listings while spot barely moved. A crowded trade clearing out its late arrivals looks alarming and signifies little about the underlying.
Why the thesis holds
The demand case predates AI and does not depend on it. Japanese restarts, life extensions in the US, Belgium and South Korea, new build across Eastern Europe and Southeast Asia, and a supply pipeline hollowed out by a decade of post Fukushima underinvestment from 2011 to 2019 that cannot be rebuilt quickly. The US alone runs an annual production gap of roughly 46 to 47 million pounds against requirements, with reshoring and strategic stockpiling now live policy. Datacentre demand was always additive to this picture, not the foundation of it.
A note on the currency. Roughly five points of the sterling line drawdown is cable strength rather than equity weakness. In dollar terms the correction is shallower than a GBP screen suggests, which is worth remembering before reading too much into the headline number.
What would break it
The thesis is not faith, and it has defined exit triggers. A sustained, multi quarter contraction in hyperscaler capex that genuinely removes the marginal demand narrative. Material unexpected supply arriving inside a three to five year window, beyond current modelling. A policy reversal driving reactor restart cancellations or capacity reductions. Or the term price breaking and holding below $80 a pound while utility contracting stalls. Absent those, a 25 to 40 percent equity correction against a flat physical market is a repricing of sentiment, not a broken thesis.
How the framework handles it
This is exactly the setup a defined framework exists for. Rather than react to the drawdown or try to guess the low, the position is scaled in tranches at pre defined depths off the highs, roughly 15, 25 and 35 percent, using the same geometry applied elsewhere in the book. On that grid the current price sits around second tranche depth. Adding at set levels rather than on conviction keeps the sizing disciplined and the decision rule mechanical, which is the entire point of having a framework when a theme is out of favour and the temptation is to either freeze or average down blindly.
The discipline is not in predicting the bottom. It is in deciding, in advance and away from the noise, what you will do at each level, and then doing only that.