The Uranium Deficit
Last month, I was doing research on AI infrastructure—specifically looking at data center economics and power requirements for a potential article on semiconductor supply chains. Pretty dry stuff, honestly.
But I kept noticing something in the earnings transcripts that didn’t add up.
Equinix mentioned their new AI-ready facilities would require 80-100 megawatts each. Digital Realty casually noted they’re seeing “unprecedented power density requirements” from hyperscale customers. Amazon’s 10-Q mentioned energy availability is becoming a “limiting factor” for AWS expansion in certain markets.
I almost skipped past it. Energy constraints aren’t exactly new—utilities always have capacity issues. But then I saw the Microsoft announcement about Three Mile Island.
Microsoft committed to a 20-year power purchase agreement to restart the Three Mile Island nuclear reactor. The one that’s been shut down since 2019. At roughly double the normal electricity rate.
That seemed... unusual.
So I started pulling threads. I spent the next week reading utility capacity reports, uranium market analysis, nuclear reactor construction timelines, and data center power consumption projections. The more I dug, the more obvious it became that something significant is happening—and most investors haven’t connected the dots yet.
Here’s what I found:
AI is creating a power crisis nobody’s really talking about.
Training a large language model like GPT-4 requires approximately 50 megawatt-hours of electricity—enough to power 50 homes for a year. Each ChatGPT query uses roughly 10x the energy of a Google search. Scale that across billions of daily queries, and the numbers get large quickly.
The data centers being built today for AI workloads require 80-100 megawatts each—comparable to a small city. Microsoft is planning 50-100 gigawatts of new data center capacity by 2030. Google, Amazon, and Meta are all on similar trajectories.
The problem is straightforward: the electrical grid wasn’t built for this.
U.S. grid infrastructure was designed around 1-2% annual demand growth. We’re now seeing 4-5% growth overall, with some regional markets (Northern Virginia, Phoenix, Dallas) projecting 10%+ annual increases concentrated in data center zones.
Dominion Energy in Northern Virginia is already telling data center developers they’re looking at 5-7 year wait times for new grid connections. Similar bottlenecks are emerging across major data center markets.
But tech companies can’t wait five years. The AI race is happening now. So they’re doing something different: going directly to power sources.
In the past 90 days:
Microsoft signed a 20-year deal to restart Three Mile Island (835 MW)
Google partnered with Kairos Power to build seven small modular reactors by 2030
Amazon invested $500 million in X-energy for SMR development
Oracle announced plans for a gigawatt-scale campus powered by three small modular reactors
Three of the world’s largest tech companies made significant nuclear commitments in under three months.
That caught my attention because of what it implies about the uranium market.
Solar and wind can’t solve this problem—they’re intermittent, and AI workloads need 99.99% uptime. Natural gas faces capacity constraints in key markets and creates ESG issues for companies with carbon-neutral commitments.
Nuclear is the only proven technology that delivers large-scale, 24/7 baseload power with zero emissions. And nuclear requires uranium.
This is where the supply-demand picture gets interesting:
Current uranium fundamentals:
Global production: ~140 million pounds annually
Current reactor demand: ~180 million pounds annually
Existing deficit: ~40 million pounds (covered by secondary sources and inventory drawdowns)
Upcoming demand additions:
60+ reactors under construction globally (China alone building 30+)
Tech companies now directly procuring nuclear power (new demand category)
Japan restarting reactors post-Fukushima (10+ already operational, more planned)
U.S. utilities extending reactor lifespans and restarting facilities
Analysts project the supply deficit reaches 60-80 million pounds annually by 2027-2028.
The constraint is timing. You can’t quickly scale uranium production. New mines require 7-10 years from discovery to first production—permitting, development, infrastructure. The supply that would meet 2027 demand needed to start development in 2017-2020.
It didn’t happen. Uranium prices collapsed after Fukushima in 2011. Mines closed, projects were shelved, exploration stopped.
Now utilities are signing long-term contracts at $80-90 per pound to secure supply—well above current spot prices around $75-80. Cameco, the largest Western producer, has publicly stated they’re essentially sold out of uncommitted production through 2027.
Uranium hit $106 per pound in February 2024—the highest since 2007. Most supply models suggest prices need to reach $120-150+ to incentivize the new production required to close the deficit.
Here’s what makes this compelling from an investment perspective:
The market is still anchored to post-Fukushima narratives about nuclear being dead. But the fundamentals have completely changed. The deficit is real, the timeline to fix it is measured in years, and utilities are already competing for scarce supply.
I’ve spent the last two weeks analyzing every publicly traded company with meaningful uranium exposure. What I found were two specific opportunities where the current valuations make no sense relative to where uranium prices are heading.
The first is a major producer trading at a significant discount to its peers—despite better margins and a structural supply advantage.
For context: Freeport-McMoRan (copper) trades at 18x forward earnings. Southern Copper trades at 22x. Both are excellent companies, but they’re mining commodities where supply can scale relatively quickly.
This uranium producer trades at just 14x forward earnings—despite:
Uranium prices at 17-year highs
The company being essentially sold out through 2027
Operating margins of 35-40% (vs. 25-30% for copper miners)
A supply deficit that’s structurally locked in for years
The valuation gap makes sense only if you believe uranium stays at $75 forever. But the supply-demand math says that’s impossible. At current production rates and demand growth, prices need to reach $120-150+ to incentivize the new supply required.
When utilities start signing 2026-2027 contracts at $90-100+ per pound in Q1 2025—which is already happening in private negotiations—this stock re-rates significantly. You want to be positioned before those numbers hit earnings calls, not after.
The second is a development-stage company with a fully-permitted project that could deliver genuinely asymmetric returns.
This company has a high-quality Kazakhstan project that could be in production by 2027. If uranium reaches $120-150 (the base case for most supply analysts I’ve read), this single project generates $400+ million in annual free cash flow at full production.
The company’s entire market cap today is $800 million.
This isn’t a lottery ticket or some junior explorer drilling holes in the desert. This is a permitted, shovel-ready project with defined resources, completed feasibility studies, and clear economics. The main risk is execution and timeline—but if they deliver, you’re looking at potential 5-10x returns over the next 3-4 years.
Why timing matters here:
Cameco—the largest publicly traded uranium producer—has already run 170% from its 2023 lows. Early investors captured that move. But based on where uranium prices are heading and the company’s forward earnings power at $120+ uranium, I think there’s another 60-80% upside from current levels.
The catalysts are approaching:
Utilities typically finalize long-term uranium contracts 12-18 months before delivery. We’re heading into the contracting season for 2026-2027 supply right now. But this time, they’re negotiating into a 60-80 million pound deficit with limited uncommitted production available.
Cameco has openly stated they have minimal uncommitted supply. Kazakhstan’s Kazatomprom—the world’s largest producer—is facing production constraints. Russian supply is politically complicated. Secondary sources (old inventories, underfeeding) are running low.
When those contract prices get announced at $90-100+ per pound—significantly above current spot prices—the market will re-rate these mining stocks accordingly. Institutional investors will update their models. Analysts will raise price targets. The sector gets a sustained bid.
But here’s what most investors miss: that re-rating happens fast. Once the numbers become obvious in earnings releases and investor presentations, the opportunity compresses rapidly.
I saw this play out with lithium miners in 2021-2022. By the time mainstream financial media was writing about the “lithium shortage,” the best-positioned miners had already doubled or tripled. The investors who made serious returns were positioned 6-12 months before the narrative became consensus.
We’re in that pre-consensus window right now with uranium.
My position:
I’ve built positions in both companies over the past two weeks. Combined, they represent roughly 8% of my total portfolio—my highest-conviction commodity allocation.
The major producer I consider relatively low-risk with 40-60% upside over the next 12-18 months. It’s profitable, generates free cash flow, has tier-1 assets, and benefits directly from higher uranium prices with minimal execution risk.
The development-stage company is higher risk—probably 40-50% downside if things go wrong (permitting delays, cost overruns, financing issues). But the upside is genuinely asymmetric. If their timeline executes and uranium reaches $120-150, this could be a 5-10x return over 3-4 years.
That risk-reward profile—60% downside, 500-1000% upside—is rare in public markets. You don’t get many chances to find permitted, shovel-ready resource projects trading at fractions of their NPV in rising commodity price environments.
What I’m watching:
To validate this thesis, I’m monitoring several specific indicators:
Near-term catalysts (Q1 2025):
Long-term contract announcements from utilities (prices and volumes)
Cameco’s Q4 2024 earnings and 2025 guidance (late January/February)
Kazakhstan production updates from Kazatomprom
Any additional tech company nuclear announcements
Medium-term indicators (2025-2026):
Spot uranium price trajectory (need sustained move above $90)
Development project timelines (particularly the Kazakhstan project I’m positioned in)
New reactor construction pace in China and India
U.S. reactor restart announcements beyond Palisades and Three Mile Island
The key question: Does the supply deficit materialize as projected, and do prices respond accordingly?
Based on everything I’ve read—utility capacity reports, uranium supply studies, reactor construction timelines—I think the answer is yes. The deficit is real, it’s growing, and there’s no quick fix.
But I’m not betting the farm on it. An 8% allocation gives me meaningful exposure if the thesis plays out, while limiting damage if I’m wrong or if execution falters.
The energy crisis driving tech companies to restart nuclear plants and sign 20-year power agreements at premium prices is real. The uranium supply-demand imbalance is structural and locked in for years. And the mining stocks haven’t fully priced this in yet.
I’ve identified two specific ways to capture this opportunity—one safer play with good upside, one higher-risk play with exceptional upside. Both are liquid, publicly traded, and positioned to benefit as this supply crunch becomes impossible to ignore.
I’m going to show you exactly which companies these are, the detailed valuation work I’ve done on each, how I’m sizing the positions, and what specific milestones I’m watching to validate or invalidate the thesis.
If you’ve been looking for a commodity play with genuine structural tailwinds and a clear 12-24 month catalyst path, this is the most compelling opportunity I’ve found this year.
Let me walk you through the full analysis.
The Two Uranium Plays I’m Building Positions In
After two weeks of digging through financials, production reports, and supply forecasts, I’ve identified two specific opportunities that give direct exposure to the uranium supply crunch without waiting for prices to fully reflect the fundamentals.
One is a large, established producer trading at a discount to its intrinsic value.
The other is a development-stage company with a world-class project that could deliver outsized returns if execution delivers.
Let me walk you through both.