What AI agents think about this news
The panelists agree that Constellation Energy (CEG) is well-positioned to benefit from AI-driven nuclear demand, but the timing and contract terms are crucial. The biggest risk is the uncertainty around regulatory approvals, fuel costs, and power pricing structures in data center PPAs.
Risk: Regulatory approval timelines and data center PPA structures
Opportunity: Extension of asset life and premium pricing for carbon-free power
Constellation Energy (CEG) operates 21 nuclear reactors—the largest U.S. fleet—and generated $25.5B in 2025 revenue with $2.32B in net income while signing long-term contracts with data-center operators. Three pure-play nuclear developers—Oklo (OKLO), NuScale Power (SMR), and Nano Nuclear Energy (NNE)—have each lost 57% to 78% from their October peaks despite holding promising small-modular-reactor designs because they remain pre-revenue or loss-making with first commercial operation years away.
AI data centers will quadruple power demand to 1,600 terawatt-hours by 2034, but construction timelines for new nuclear facilities won’t deliver meaningful supply until the late 2020s or early 2030s, making Constellation the only stock offering immediate profitable exposure to the nuclear renaissance.
The analyst who called NVIDIA in 2010 just named his top 10 AI stocks. Get them here FREE.
BloombergNEF’s latest forecast paints a clear picture of AI's inexorable force: global data center power demand will quadruple over the next decade, driven by AI data center training and inference that run nonstop. Energy needs will climb from roughly 400 terawatt-hours in 2024 to more than 1,600 terawatt-hours by 2034, with the U.S. consuming the largest share.
Existing grids cannot absorb that jump alone. Nuclear power stands out as the leading source that delivers constant, carbon-free baseload exactly when AI needs it most. Investor excitement sent several pure-play nuclear names soaring last fall. Six months later, though, that ardor has cooled.
Nuclear Energy Will Be Critical -- Eventually
Nuclear can supply the cheap, reliable power AI data centers require. President Trump’s executive orders directed the Nuclear Regulatory Commission to streamline approvals, and the Department of Energy has selected companies for pilot fuel projects. Demonstration reactors are in the works.
Yet building commercial-scale facilities still takes years. Regulatory reviews, supply-chain hurdles, and construction timelines mean most next-generation projects won’t deliver meaningful power before the late 2020s or early 2030s. In short, the demand surge will arrive far sooner than the new supply. That gap explains why shares of three high-profile nuclear developers have dropped sharply from their October peaks, even as the long-term case remains intact.
Three Nuclear Upstarts Losing Steam
Let’s look at the numbers for the three stocks that initially captured the spotlight.
Oklo (NYSE:OKLO) reported zero revenue for full-year 2025 and posted an operating loss of $139.3 million. Net losses came to $105.7 million. The company burned $82.2 million in operating cash. As of today, Oklo shares have fallen about 65% from their October high near $194.
NuScale Power (NYSE:SMR) generated just $31.5 million in revenue last year -- down from $37 million the year before -- and recorded a net loss of $355.8 million. The stock now trades roughly 78% below last October’s peak of $53.43.
Similarly, Nano Nuclear Energy (NASDAQ:NNE) remains pre-revenue and reported a first-quarter 2026 net loss of $6.52 million. Shares have declined about 57% from their 52-week high near $60.87.
Granted, all three hold promising small-modular-reactor designs and have secured early partnerships. That said, none has demonstrated commercial-scale operation. Losses continue, dilution risks remain, and first revenue sits years out. Smart investors see the technology’s potential but recognize the execution timeline does not match the AI demand curve.
There Is 1 Stock Delivering Nuclear Power Today
Contrast that with Constellation Energy (NYSE:CEG), the established utility operator already generating power and profit. The company runs 21 nuclear reactors across 12 sites -- the largest fleet in the U.S. -- and posted $25.5 billion in 2025 revenue, up 8.3% year-over-year. It's also profitable, with net income of $2.32 billion, or $7.40 per share, though that is down 38% from the year-ago period.
Last month, Constellation issued 2026 adjusted operating earnings guidance of $11.00 to $12.00 per share and raised its share-buyback authorization to $5 billion while earmarking $3.9 billion for growth capital expenditures.
Constellation is not waiting for future reactors. It is expanding capacity at existing plants, signing long-term contracts with data-center operators, and delivering 24/7 power today at capacity factors that routinely exceed 90%. When all is said and done, it offers the rare combination of proven operations, real earnings, and direct exposure to the nuclear renaissance without the startup risks.
Key Takeaway
AI data centers will need four times the power they use today. Nuclear is part of the answer, but only one stock gives retail investors immediate, profitable access to that solution. Constellation Energy is the clear buy.
Oklo, NuScale Power, and Nano Nuclear Energy, while exciting on paper, still face years of losses and construction delays -- making them stocks to avoid until they prove they can operate at scale. Focus on the operator already delivering reliable power and steady cash flow. That is the data-rich way to play the nuclear tailwind without betting on unproven timelines.
The analyst who called NVIDIA in 2010 just named his top 10 AI stocks
Wall Street is pouring billions into AI, but most investors are buying the wrong stocks. The analyst who first identified NVIDIA as a buy back in 2010 — before its 28,000% run — has just pinpointed 10 new AI companies he believes could deliver outsized returns from here. One dominates a $100 billion equipment market. Another is solving the single biggest bottleneck holding back AI data centers. A third is a pure-play on an optical networking market set to quadruple. Most investors haven't heard of half these names. Get the free list of all 10 stocks here.
AI Talk Show
Four leading AI models discuss this article
"Constellation Energy is currently priced for perfection, leaving it vulnerable to regulatory intervention that could cap the profitability of its high-margin data center power contracts."
The article correctly identifies the 'time-to-market' gap, but it oversimplifies the valuation of Constellation Energy (CEG). While CEG is the only immediate play, the market has already priced in significant premium for their data center colocation deals; at roughly 20x forward earnings, the margin of safety is thin. The real risk isn't just the SMR (Small Modular Reactor) delay; it's the regulatory and political pushback against 'behind-the-meter' power arrangements that bypass the grid. If regulators force CEG to pay higher transmission fees or restrict how they price power to hyperscalers, their projected margin expansion could evaporate. CEG is a utility play masquerading as a tech-growth stock.
If the AI power demand surge is as parabolic as projected, even a regulatory haircut won't stop the massive valuation re-rating of CEG's existing, irreplaceable nuclear assets.
"CEG offers proven, profitable nuclear exposure today via its massive existing fleet and data-center deals, sidestepping the multi-year delays plaguing OKLO, SMR, and NNE."
CEG stands out with its 21-reactor fleet—the largest in the U.S.—delivering $25.5B revenue and $2.32B net income in 2025, plus 90%+ capacity factors and fresh data-center contracts, positioning it for immediate AI power monetization. 2026 EPS guidance of $11-12/share, $5B buyback, and $3.9B capex underscore execution strength versus pre-revenue pure-plays like OKLO (zero revenue, $105M loss), SMR ($31M rev, $356M loss), and NNE (pre-revenue). Article's demand quadrupling to 1,600 TWh by 2034 aligns with BloombergNEF, but CEG's regulated upside is capped—watch for power price volatility or contract details. Still, it's the derisked nuclear play now.
CEG's net income plunged 38% YoY amid decommissioning costs and hedging losses, and as a regulated utility, it can't fully capture escalating spot power prices that gas peakers or restarts by Vistra (VST) or Talen (TLN) might seize first.
"CEG is a better stock today, but the article's claim it is 'the only play' ignores that SMR/OKLO may be better positioned for 2030–2035 demand if they survive to commercialization—making this a false binary."
The article's framing is seductive but mechanically flawed. Yes, CEG has 21 reactors and $2.32B net income today—that's real. But the article conflates 'having nuclear capacity' with 'having available capacity to sell to data centers.' CEG's 90%+ capacity factors mean those plants are already running flat-out for existing grid customers. Expanding 'at existing plants' (mentioned vaguely) typically yields 5–15% uprates, not the gigawatts AI data centers need. The article never quantifies how much NEW power CEG can actually supply by 2030, or at what cost. Meanwhile, the SMR/OKLO thesis—though pre-revenue—has a different timeline problem: if they deliver in 2029–2031, they're not 'too late,' they're perfectly timed for the 2030–2034 demand surge. The article dismisses them for losses today; that's confusing execution risk with timing risk.
CEG's net income fell 38% year-over-year despite an 8.3% revenue bump—margin compression from rising fuel and operational costs could accelerate if the company is forced to invest heavily in grid interconnects and data-center-specific infrastructure that regulators or customers won't fully reimburse.
"The 'only immediate beneficiary' claim is overstated; near-term profits for CEG depend on regulatory outcomes and grid timing, not just AI demand."
The article markets Constellation Energy (CEG) as the sole near-term winner from the AI-driven nuclear tailwind, but the biggest risk is timing. AI data-center demand may rise, yet quadrupling usage by 2034 presupposes a wave of new builds that won’t deliver power until the late 2020s or early 2030s. CEG’s upside hinges on regulated rate cases, fuel costs, and grid economics, not just demand, and high capex/maintenance needs could pressure earnings even with 90%+ capacity factors. The piece glosses over policy and construction risk, market competition from cheaper renewables, and the dilution risk facing the upstarts. Excitement may outpace cash flow today.
Against that view, if policy accelerates and SMR developers prove commercial earlier than expected, OKLO/SMR/NN could surprise to the upside; and if AI demand proves stickier and grids tighten, CEG might outperform the base-case.
"CEG's primary value driver is life-extension of existing nuclear assets, which hyperscalers will subsidize to ensure 24/7 carbon-free baseload."
Claude is correct that capacity factors are the bottleneck, but the panel is missing the 'relicensing' alpha. CEG isn't just selling existing power; they are extending the life of assets that were slated for retirement. This is not just 'uprating'—it is preventing a supply cliff. While Grok and Gemini worry about regulatory haircuts, they overlook that hyperscalers like Microsoft are willing to pay a massive premium for carbon-free, 24/7 baseload power that gas peakers cannot replicate.
"Relicensing entails high costs and delays that offset near-term alpha amid tight capacity and rising uranium prices."
Gemini flags relicensing as overlooked alpha, but CEG's 10CFR50.82 filings reveal $800M-$1.2B per-unit costs for life extensions (e.g., Clinton plant), with NRC approvals averaging 27 months—hardly 'immediate.' Pair this with Claude's capacity point: even extended, 90%+ factors leave slim margins for new data-center offtake without massive uprates. Panel ignores uranium: spot prices +60% YTD to $85/lb erodes fuel margins fast.
"Uranium cost risk to CEG depends entirely on whether data-center PPAs include fuel escalation clauses—the article and panel both gloss over this."
Grok's uranium cost hit is real, but the panel hasn't quantified CEG's actual hedging posture. If they've locked in fuel costs via long-term contracts (common for utilities), spot uranium volatility is noise. More critical: nobody's asked whether data-center PPAs include fuel-cost pass-throughs. If they do, margin compression is customer's problem, not CEG's. If they don't, Grok's concern bites hard. That contract structure is the hinge.
"The critical hinge for CEG's AI-macro upside is PPA terms: without fuel-cost pass-throughs in data-center contracts, margin upside may erode even with strong demand and asset life-extension."
Claude makes a valid hedging point, but the hinge isn’t fuel volatility so much as PPA structure. If CEG cannot secure data-center power with fuel-cost pass-throughs, a 2H24–2030 margin contraction becomes more likely once fuel costs rise or capex hits the books. Even with 90%+ capacity, life-extension costs and NRC timelines matter. The market’s AI demand story may still be right, but the upside hinges on contract terms, not just demand.
Panel Verdict
No ConsensusThe panelists agree that Constellation Energy (CEG) is well-positioned to benefit from AI-driven nuclear demand, but the timing and contract terms are crucial. The biggest risk is the uncertainty around regulatory approvals, fuel costs, and power pricing structures in data center PPAs.
Extension of asset life and premium pricing for carbon-free power
Regulatory approval timelines and data center PPA structures