By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
While AI-driven capex increases from hyperscalers signal long-term demand for nuclear power, near-term benefits for uranium producers like Cameco are debated due to construction timelines, regulatory risks, and competition from established producers. Hyperscalers' direct PPAs with utilities may accelerate uranium demand, but this is not yet widely agreed upon.
Risk: Secondary market risk and stranded power risk, as highlighted by Gemini, with potential regulatory intervention or windfall taxes on utilities.
Opportunity: Accelerated uranium demand due to hyperscalers' direct PPAs with utilities, as suggested by Grok.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
If there was any worry that spending on artificial intelligence (AI) was finally set to slow down, it's just been wiped away. Not only are big tech companies not dialing back their aggressive investments in AI infrastructure, but they're ramping them up.
Google parent Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL) recently raised its 2026 capital expenditure (capex) forecast from a range of $180 billion to $190 billion to a range of between $195 billion and $205 billion. Amazon (NASDAQ: AMZN) upped its capex outlook for this year from a previous estimate of $200 billion to a new estimate of $220 billion. Facebook parent Meta (NASDAQ: META) is putting more money into AI as well, raising this money by issuing new debt.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
These expanded spending plans obviously bode well for a company like Nvidia (NASDAQ: NVDA), which still makes the majority of the processors at work within AI data centers. It's also a boon for less obvious, indirect beneficiaries such as Vertiv (NYSE: VRT) or GE Vernova (NYSE: GEV). The former offers power-distribution and chip-cooling solutions, while the latter makes electricity-generating natural gas turbines that can power an entire AI data center.
Perhaps the most underappreciated and undervalued company that wins from these growing capital expenditure budgets, however, is Cameco (NYSE: CCJ). It provides much of the uranium needed by nuclear power plants, which are increasingly important sources of the electricity that AI data centers desperately need.
Cameco is in the right place at the right time
It's not the only name in the uranium business. It's technically not even the biggest name in the business. That title belongs to Kazakhstan's Kazatomprom, which serves many of its geographic neighbors to account for about a fifth of the industry's total global supply.
Saskatchewan, Canada-based Cameco is a major player, though, and certainly critical to North America's nuclear power industry. It sold 33 million pounds of triuranium octoxide last year, which you may know better as the yellowcake uranium fuel required by most nuclear reactors.
It's not just a miner or refiner, though. The company also owns a 49% stake in Westinghouse Electric, which makes and services nuclear power reactors all over the world. Its tech is the heart of more than 90 facilities in 21 countries, in fact, providing Cameco with another (albeit indirect) foothold in the business. Last year, the company turned revenue of $3.5 billion into adjusted net earnings of $627 million, well up from 2024's figures.
And business is about to be even better.
With the planet struggling to keep up with ever-growing demand for electricity while also working to dial back its dependence on polluting fossil fuels, well-proven nuclear power is back in vogue. The World Nuclear Association believes worldwide nuclear power capacity could easily double between now and 2050, with some projections suggesting it could triple during this time. In this vein, Goldman Sachs predicts the current global count of about 440 reactors could reach 500 as soon as 2030, although another 400-plus are either already proposed or planned, according to the World Nuclear Association.
Their one common thread? Almost all of them will be powered by uranium-235, a great deal of which starts its journey at one of Cameco's mines.
Think bigger picture and longer term
But is the company actually going to benefit from the artificial intelligence spending frenzy that's underway right now?
If it does, it probably won't happen overnight. It takes years to build and activate a new nuclear power facility, but AI data centers need more electricity right now. A company like the aforementioned GE Vernova is seemingly better positioned to meet the more immediate need.
Just think bigger picture and longer term. GE Vernova's natural gas power turbines are still mostly a stopgap solution. Indeed, the ramped-up consumption of natural gas they're causing could also pump up the price of the commodity itself, ultimately making this particular on-premise option a less cost-effective one. Although it costs more up front, in the long run, nuclear power is achieving, or even exceeding, cost parity with alternatives.
And as it turns out, many nuclear power plants built decades ago are still functioning just fine, outlasting initial estimates of how long they'd remain viable. The U.S. Nuclear Regulatory Commission, or NRC, reports that some facilities initially licensed for up to 40 years could end up lasting 80. And that's with an older design. Newer designs and materials could have even longer lifespans than the facilities being built and planned today.
Then there's another development: Technology companies and public utilities are evolving beyond the typical customer/provider relationship to cultivate true partnerships. As an example, in late 2024, software giant Microsoft and power utility provider (and major nuclear power plant operator) Constellation Energy (NASDAQ: CEG) co-announced plans to restart one of the idle nuclear reactors at Pennsylvania's Three Mile Island to provide power for one of Microsoft's data centers.
Utility outfit Vistra (NYSE: VST) is also working directly with Amazon and Meta to provide nuclear power to some of their AI data centers, justifying Vistra's investment in such facilities that will connect directly to the same power grid that serves retail customers.
If nothing else, it's a testament to the flexibility of nuclear power and the possibility of prolonging or even expanding the power output of an existing nuclear power facility. It's not like artificial data centers or the rest of the world are going to need less power in the future.
Just keep things in perspective
The point is, although the knee-jerk bullishness that sent Cameco shares soaring last year on the heels of AI's explosion seems to have run its course (the stock's been lackluster so far this year), that's not necessarily the end of the story. That's just the first of several chapters.
While the second act may not be nearly as exciting as the first one -- and will take even longer to play out -- it will be the one that separates the winners from the losers, based on profits or lack thereof. Cameco will almost certainly come out of this period as a winner, particularly once the nuclear power plants currently under construction go online and begin consuming enriched uranium.
Despite Cameco's lethargic share performance this year, most members of the analyst community still rate this stock a strong buy, with an average target of $125.25, almost 30% more than now. That's not a bad short-term tailwind to start a new long-term position with.
Should you buy stock in Cameco right now?
Before you buy stock in Cameco, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Cameco wasn't one of them. The 10 stocks that made the cut are built for long-term growth and could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $421,511!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,381,960!*
That performance is why people listen. With a track record of beating the S&P 500 by nearly 5x, Stock Advisor offers a distinct advantage. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built for the long haul.
James Brumley has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Amazon, Cameco, Constellation Energy, GE Vernova, Goldman Sachs Group, Meta Platforms, Microsoft, Nvidia, Vertiv, and Vistra. The Motley Fool has a disclosure policy.
Four leading AI models discuss this article
"Cameco wins in a multi-year nuclear buildout scenario but near-term AI power needs favor quicker alternatives, and the article overstates current momentum while understating execution risks."
The article correctly notes that Alphabet, Amazon, and Meta's upward capex revisions (GOOG to $195-205B by 2026, AMZN to $220B this year) signal sustained AI infrastructure buildout, benefiting uranium via nuclear power demand. Cameco (CCJ) benefits long-term as a major North American uranium producer with Westinghouse exposure, amid projections of global nuclear capacity doubling by 2050. However, the piece glosses over near-term nuclear construction timelines (5-10+ years), permitting risks, and competition from Kazatomprom (20% global supply). CCJ's 2024 revenue was actually ~$2.5B, not $3.5B, and its stock is up ~35% YTD, not lethargic.
Nuclear renaissance timelines are routinely delayed by regulatory hurdles, cost overruns, and public opposition; AI data centers may increasingly turn to faster-deployed gas, renewables plus storage, or even small modular reactors from competitors, leaving Cameco's uranium uplift years away and vulnerable to oversupply if Kazakhstan ramps production.
"The transition from speculative AI software to physical infrastructure requires a guaranteed, 24/7 baseload power source that only nuclear can provide at scale, making uranium miners essential, long-term utility-like assets."
The massive capex ramp from Hyperscalers confirms a structural shift in power demand, moving beyond mere 'AI hype' into tangible, long-term infrastructure requirements. While the article correctly identifies nuclear as a beneficiary, it underestimates the regulatory and logistical friction. Cameco (CCJ) is a primary play, but the market is pricing in a 'nuclear renaissance' that faces severe supply chain bottlenecks in enrichment and reactor construction timelines. I am bullish on the sector, but investors must look past the spot price of uranium and focus on companies with existing, permitted capacity that can bypass the decade-long lead times for new builds. The real alpha is in the grid-level power providers like Constellation Energy (CEG) who control the existing, baseload assets.
If the 'AI energy' thesis is wrong and LLM efficiency gains outpace compute demand, or if regulatory hurdles permanently stall nuclear restarts, these high-capex utilities will face a massive stranded asset problem.
"Cameco is a genuine long-term nuclear play, but the current capex cycle benefits near-term power infrastructure names (GEV, VRT) far more than uranium miners in the next 2-3 years."
The article conflates two different timelines and conflates them into one thesis. Yes, capex is rising—that's real. Yes, nuclear is having a moment—also real. But the article glosses over a critical gap: AI data centers need power *now*, not in 2030-2035 when new reactors come online. Cameco benefits from a 10-15 year structural tailwind in nuclear, but that's decoupled from the current capex surge. The stock's underperformance YTD despite rising uranium demand suggests the market already priced in the long-term thesis. The immediate beneficiaries are GE Vernova (gas turbines) and Vertiv (cooling)—not Cameco. Uranium is also cyclical and geopolitically volatile; Kazakhstan supplies 20% globally, and Kazatomprom (not listed in NA) is the largest producer.
If the article is right that utilities and tech are forming direct partnerships (Microsoft-Constellation, Amazon-Vistra), then existing reactor restarts and life extensions could accelerate uranium demand faster than the article's 2030+ timeline suggests, pulling forward Cameco's upside.
"AI infrastructure capex is a structural tailwind, but its profitability depends on energy costs and project execution; without favorable energy and financing conditions, the upside may be overstated."
Even with big capex bumps at Alphabet, Amazon, and Meta, the takeaway is not a guaranteed profits surge. The article treats capital expenditure as a one-way bet on AI infrastructure, but capex is highly cyclical and often front-loaded, with returns contingent on energy costs, grid reliability, and financing terms. The nuclear angle—Cameco and uranium demand—feels speculative: reactor buildouts are lengthy, regulatory risk is real, and demand for yellowcake hinges on many moving parts beyond AI data centers. Moreover, hyperscalers can re-optimize workloads or delay capacity on a cooling or power price shock. If energy costs rise or rates stay high, the ROI story weakens.
The strongest counter-argument is that this feels priced into the shares; unless energy costs stay favorable and financing remains cheap, realized returns may disappoint. Additionally, nuclear optimism may be overdone, given regulatory delays and long construction timelines.
"Existing reactor restarts and direct hyperscaler PPAs accelerate Cameco's uranium uplift well before 2030."
Claude correctly flags the now-vs-2030 power gap, yet overlooks that Microsoft-Constellation and Amazon-Talen deals already restart reactors and sign PPAs pulling uranium forward. CCJ's 2025-2027 delivery books tighten faster than spot suggests; dismissing near-term demand underestimates how hyperscalers are bypassing new-build timelines entirely.
"Regulatory intervention and windfall taxes on hyperscaler-linked power deals pose a greater risk to equity valuations than the supply-side bottlenecks currently discussed."
Claude and Grok are missing the secondary market risk: the 'direct purchase' model between hyperscalers and utilities like Vistra (VST) or Constellation (CEG) effectively ring-fences power supply, creating a two-tier market. This creates a massive 'stranded power' risk for local industrial users, likely triggering regulatory intervention or windfall taxes on these utilities. If governments cap energy prices to protect the grid, the margin expansion that investors are currently pricing into these nuclear-linked equities will evaporate overnight.
"Reactor restarts pull forward power demand, not uranium supply—a critical distinction the panel is blurring."
Gemini's two-tier market risk is real, but the windfall tax threat is speculative—no major jurisdiction has signaled this yet. More pressing: hyperscalers' direct PPAs with utilities (Microsoft-Constellation, Amazon-Talen) lock in *existing* reactor capacity, not new uranium demand. Grok conflates PPA acceleration with yellowcake tightening. Cameco's near-term uplift depends on *new* enrichment and mining capacity, not just reactor restarts. The article's uranium thesis remains 2030+ dependent.
"Near-term Cameco upside depends more on new supply and permitting than on hyperscaler PPAs alone, even as two-tier market risks exist."
Gemini’s two-tier market risk is real but not the dominant near-term driver. The more pressing risk is whether financing and permitting for new reactors materialize quickly enough to pull uranium demand forward, or whether policy shocks like windfall taxes erode utility margins and dampen PPAs. Cameco’s upside still hinges on new supply, not just PPAs. If financing conditions stay loose and projects clear permitting, uranium demand could surprise to the upside.
While AI-driven capex increases from hyperscalers signal long-term demand for nuclear power, near-term benefits for uranium producers like Cameco are debated due to construction timelines, regulatory risks, and competition from established producers. Hyperscalers' direct PPAs with utilities may accelerate uranium demand, but this is not yet widely agreed upon.
Accelerated uranium demand due to hyperscalers' direct PPAs with utilities, as suggested by Grok.
Secondary market risk and stranded power risk, as highlighted by Gemini, with potential regulatory intervention or windfall taxes on utilities.