When Democrats secured Congress in the midterms, investors wasted no time dumping shares across the nuclear and AI power sector for a second consecutive session on November 5.
Oklo (NASDAQ: OKLO) shed 13% in a single day, while Bloom Energy cratered 16%, as markets repriced the assumption that a hyperscaler-driven electricity boom would underpin the entire sector.
The selloff exposed a fundamental question: how much of the nuclear investment case actually depends on data centers, and how much was always about something else entirely?
Cameco (NYSE: CCJ) CEO Timothy Gitzel addressed that question directly on his company’s third-quarter earnings call, drawing a clear line between the AI narrative and his firm’s core business model.
“Artificial intelligence and data centers can provide meaningful upside, but they are not the foundation of the nuclear investment thesis nor are they a load-bearing assumption in our strategy,” Gitzel said.
His argument rests on contracted fuel demand from reactors already generating power today, a base that continues burning uranium regardless of what hyperscalers decide to build next quarter.
Cameco had 230 million pounds of uranium committed across 39 customers at the close of 2025, providing revenue visibility that few companies in the sector can match.
Analysts project the company’s revenue will climb from C$3.48 billion in 2025 to C$4.34 billion by 2028, a roughly 25% increase driven primarily by those existing fuel contracts rather than any speculative AI demand.
Even so, Cameco’s forward price-to-earnings multiple collapsed from 111x in April to 47x, now sitting below its three-year average of 61x, suggesting the market had already priced in a substantial AI premium that is now deflating.
Gitzel acknowledged the fragility of theme-driven investing himself, noting that “capital can easily rotate out of a theme in a day or in a week,” a warning that proved prescient given the November selloff.
NuScale Power (NYSE: SMR) finds itself in a structurally different position, with CEO John Hopkins pushing back against the data-center characterization on the company’s third-quarter call the same day markets sold off.
Hopkins said it was “probably too early to tell” where data center demand ultimately settles, adding the critical qualifier that the company is “not reliant on data centers.”
That claim is technically accurate, though for a less reassuring reason: NuScale generated just $0.3 million in third-quarter revenue, down sharply from $8.2 million in the same period a year earlier.
The company’s growth trajectory depends almost entirely on its partner ENTRA1 finalizing a power purchase agreement with the Tennessee Valley Authority, a step that has not yet been completed.
Hopkins indicated that once the TVA agreement is signed, NuScale will proceed immediately into licensing work and procurement of long-lead components, though management clarified that the PPA itself does not constitute a module sale.
A separate OEM agreement between NuScale and ENTRA1 would need to follow before any actual module transaction occurs, adding another layer of dependency to an already thin revenue base.
Analysts nonetheless project a dramatic revenue acceleration, from $20.4 million in 2026 to $127 million in 2027 and $236 million in 2028, a ramp that requires the TVA process to stay on schedule.
NuScale ended September holding $1.9 billion in cash and investments, giving it a financial runway that reduces near-term default risk even as revenue remains negligible.
On the political dimension, Hopkins expressed confidence that the midterm outcome would not materially alter the nuclear landscape, pointing out that “both the Ds and the Rs are pro-nuclear.”
That assessment aligns with a Pew Research Center survey from 2025 showing 59% of Americans supporting construction of more nuclear plants, up from just 43% in 2020, reflecting a genuine shift in public sentiment rather than a partisan position.
Oklo presents the starkest risk profile among the group, with analysts forecasting per-share losses of $0.98 this year and $1.06 next, and its business model far more directly tied to the data-center buildout than either Cameco or NuScale.
The central irony of the November selloff is that investors treated Cameco as though its fortunes rise and fall with hyperscaler capital expenditure budgets, when its contracted uranium book tells a considerably more stable story.
