This group sells the raw fuel that powers nuclear reactors: uranium, dug out of the ground and processed into a yellow-brown powder called yellowcake (chemically U3O8, "triuranium octoxide"). A reactor cannot run without it, and nothing else can substitute. The companies here are miners and fuel processors — Cameco (CCJ), Uranium Energy (UEC), Denison (DNN), Energy Fuels (UUUU), Centrus Energy (LEU), and NexGen (NXE). Their customers are electric utilities that own nuclear power plants, and increasingly the data-center operators who want round-the-clock carbon-free electricity to run AI compute. The reason this matters now: AI/AGI data centers need large amounts of firm, always-on power, and nuclear is one of the cleanest large-scale sources of it — so every reactor that gets built, restarted, or run harder pulls more uranium out of a supply chain that takes years to expand.
In one breath: forward demand looks set to exceed forward supply on the figures here, because reactors are being built, restarted, and run harder while new mines take roughly 7-10 years to come online and a depleting pile of old stockpiles is what currently plugs the gap. The thing that limits supply is not running out of uranium in the earth's crust — it is permitting, capital, lead time, and the fact that one country and a handful of companies control most of the output. In money terms, after a multi-year price run the listed miners trade at high multiples of today's earnings (you pay a lot per dollar of current cash they produce), and most of the smaller names are not yet producing real cash at all — so much of what you buy is a claim on future tonnes, not present profit. This is a description, not a recommendation; the reader judges.
The product is uranium concentrate — U3O8, "yellowcake." It is mined three ways: open-pit (digging a big hole), underground, or in-situ leach / ISL (pumping fluid down a well to dissolve uranium in place and pumping it back up — generally cheaper and lower surface impact). The raw concentrate then passes through a long fuel chain before it can power a reactor: conversion (turning the powder into a gas), enrichment (raising the concentration of the fissile isotope U-235, the splittable atom that sustains the reaction), and fabrication (pressing it into fuel pellets and rods). The miners in this group mostly sell at the first step; a couple reach further down the chain.
The unit of sale is the pound of U3O8, priced in US dollars per pound. Worth understanding clearly: uranium is a small share of the total cost of running a nuclear plant (the plant itself, financing, and staffing dwarf it), but it is 100% essential and has no substitute — which is why utilities care far more about securing supply than about shaving the fuel price. That shapes how the money is made:
Cash mechanics for an owner: a miner spends heavily up front to build a mine (capex = capital expenditure, the money sunk into building productive assets), then earns a margin on every pound for years. The key swing factor is the gap between the price it sells at and its cash cost to pull a pound out of the ground. When the uranium price is well above cash cost, mature producers generate real owner cash; when it isn't, they lose money or idle the mine and wait.
Today (known/steady-state): the world's reactor fleet consumes roughly 180 million pounds of U3O8 per year. est. That is a relatively steady, non-discretionary number — a running reactor needs its fuel reload on schedule regardless of the economy. Buyers are electric utilities and, increasingly, the fuel-cycle intermediaries (traders and converters between miner and utility) and governments rebuilding strategic stockpiles.
Forward (forecast): demand is set to climb from several stacking drivers, all of which the AGI lens amplifies:
In dollar terms, even at a steady ~180M lbs/yr the market is on the order of est. low-tens-of-billions of dollars a year at recent prices (180M lbs × roughly the prevailing per-pound price); the forward path is up on volume as fleets grow. Note this is a much smaller dollar market than, say, oil or gas — uranium's importance is its irreplaceability, not its dollar size.
✓ VERIFIED — the following figures were confirmed from primary sources after initial publication:
The shape of the problem: new mine production currently covers only roughly 70-80% of reactor demand. est. The remaining slice has been filled by "secondary supply" — government and commercial stockpiles, plus enrichment techniques like underfeeding (running enrichment plants in a mode that produces extra fuel from existing material instead of fresh uranium) — and those secondary sources are being drawn down. As the buffer depletes, primary mines have to cover more of the load, but they can't expand quickly.
The main bottleneck is not geology — it is time, permits, and capital. There is plenty of uranium in the earth's crust; the constraint is that a new mine takes roughly 7-10 years to permit and build, requires large up-front capital, and depends on specialised know-how. est. That long lead time is among the most important facts in this whole group: it means supply cannot respond fast even when prices rise, so a demand surge can keep the market short for years. A secondary bottleneck sits further down the chain — enrichment capacity, especially for HALEU, where the West is thin and Centrus (LEU) holds the only US license. New reactors can be slowed by fuel-cycle gaps even when raw uranium exists.
Who controls supply (market structure):
Put plainly: on the figures in the scan, the product is short — described there as structural, not just momentary. Mines make roughly 70-80% of what reactors burn; the rest comes from a shrinking stockpile; demand is rising from restarts, new builds, and AI-driven baseload appetite; and new supply takes the better part of a decade to arrive. The evidence cited for the deficit being structural rather than theoretical: the multi-year drawdown of secondary supplies, utilities returning to long-term contracting to lock in pounds, and the scan's own framing of the gap as "structural and widening."
| Factor | Now | Forward direction | What it means for the gap |
|---|---|---|---|
| Reactor demand | ~180M lbs/yr est. | Rising (restarts, new builds, AI baseload) | Widens the gap |
| Primary mine supply | ~70-80% of demand est. | Rises slowly (7-10 yr lead time) | Can't close gap quickly |
| Secondary supply (stockpiles) | Filling the remaining ~20-30% est. | Depleting | Removes the cushion → gap bites |
| Supply concentration | ~40% Kazakhstan est. | Geopolitically fragile | Raises odds of supply shock |
| Enrichment / HALEU | Thin in the West | Expanding from a low base | Separate bottleneck for new reactors |
Source: 500-stocks scan, /Users/ravf/projects/work/.claude/worktrees/sector-hub/research/investments/500-stocks/01-energy-power.html (Sector 03, "Nuclear — Uranium Miners"). Numeric shares/lead-times are approximate and not live-verified.
When could it flip to oversupply? A deficit driven by long mine lead times does not reverse fast — but it is not permanent either. The conditions that would close or invert the gap, none of which are imminent on the scan's facts: (a) the wave of new mines now in permitting/construction actually delivering on time over the late-2020s; (b) Kazakh and other low-cost output ramping hard into high prices; (c) reactor restarts and SMR build-out disappointing relative to the bullish forecasts; (d) a global recession or AI-capex pullback softening power demand. Because supply takes years to respond, the realistic window for any flip to oversupply is a multi-year-out question, and it depends heavily on whether the AGI-driven demand forecast materialises as booked reactors. forecast
The group splits into one large diversified producer, a set of US/Canadian developers and small producers, and one fuel-cycle (enrichment) specialist. Exposure to "this product" — raw uranium pounds — varies a lot. The size figures below are general-knowledge approximations and not live-verified.
| Ticker | What it makes | Exposure to uranium | Rough size est. | Position / edge |
|---|---|---|---|---|
| CCJ (Cameco) | Mined uranium + stake in fuel/services (incl. Westinghouse interest) | High; the most diversified-but-pure of the group | Large-cap (tens of $B) est. | Largest Western producer; most liquid; tier-1 Canadian assets |
| UEC (Uranium Energy) | US in-situ leach uranium | Pure-play, US-focused | Small/mid-cap est. | Permitted US ISL capacity ready to restart; "Western supply" premium |
| NXE (NexGen) | Development-stage Canadian uranium (Arrow deposit) | Pure-play, pre-production | Mid-cap est. | Very large high-grade undeveloped deposit; claim on future tonnes, not current cash |
| DNN (Denison) | Development-stage Canadian uranium + physical uranium holdings | Pure-play, mostly pre-production | Small/mid-cap est. | Athabasca Basin development; holds physical uranium as a balance-sheet asset |
| UUUU (Energy Fuels) | Uranium and rare-earth elements (plus vanadium) | Diversified — uranium is one of several revenue lines | Small/mid-cap est. | US conventional mill (White Mesa); rare-earth angle adds a separate critical-minerals line |
| LEU (Centrus Energy) | Uranium enrichment & fuel services (incl. HALEU) | Down-chain, not a miner — exposure is to fuel demand, not pounds in the ground | Small/mid-cap est. | Only US-licensed HALEU enricher; strategic for SMRs and US fuel independence |
Source: company identities and positioning from the 500-stocks scan (Sector 03); sizes and segment mixes are general-knowledge approximations, not live-verified. est.
Reading this table for "who captures the money": CCJ is the one currently producing at scale and earning real cash; UUUU earns cash but uranium is only part of it; UEC sits on producible capacity; NXE and DNN are largely claims on future production; LEU captures money from a different point in the chain (enrichment) and is exposed to the HALEU/SMR fuel bottleneck rather than to raw pounds. This describes where each sits in the chain, not which to own.
Here is the plain-money translation, kept to neutral arithmetic. After a multi-year run in both uranium prices and these equities, the listed miners generally trade at high multiples of their current earnings and cash flow — meaning you pay a large amount of market value today for each dollar of profit they are making right now. est. Mechanically, that is what it looks like when a market is pricing in a future deficit rather than present results: much of the share price reflects pounds and prices that haven't been booked yet.
Money-in / money-out shape of the group: this is a high-capex, capital-intensive business, not a capital-light one. Building and sustaining mines absorbs large cash before it returns any; the developers are net cash-consumers today (raising money, then spending it on permitting and construction) and the producers swing between strong cash generation when the uranium price sits well above their cash cost and thin or negative cash when it doesn't. So as an owner you are buying into a cyclical, lumpy cash profile where the central variable is the per-pound price relative to cost, and you are currently paying full or premium multiples for that exposure on general-knowledge figures. How that price compares to value depends on how high and how durable the future uranium price proves; this section states the arithmetic and leaves the judgment to the reader.
Where a company-level deep-dive would add the most information, by role in the group (this is a map of where information sits, not a ranking of what to buy):
Used:
Hard vs approximate (be explicit):
Source: as listed above. Live web retrieval was unavailable at build time; verify all approximate figures against current filings and live pricing before acting on them.