RAW SCIENCE

Series: Critical Analysis of Uranium: The Fuel-Cycle Squeeze, Part 1 of 4.

Table of Content

Table Of Contents

1. Uranium Enrichment… Not Only the Mine 2. The four steps of uranium 3. Why can't you just dig up reactor fuel? 4. SWU: the separation and enrichment unit 5. The dial: feed versus work * Let´s see it in numbers * Continue reading with Pro * Let´s see it in numbers 6. The geopolitical situation of uranium enrichment: will we need more or less mining? * The decades of underfeeding and Russian cheap enrichment 7. Russia's quiet centrality and the 2024 policy shock * The Ukraine war and the consequences of uranium enrichment 8. The chain that almost nobody draws * Sizing the event. 9. What would break this thesis? 10. Editorial Note * Methodology appendix: feed factor and SWU calculations 11. References

Uranium Enrichment… Not Only the Mine

Uranium is not like copper or silver. For most metals, extraction is the key stage: the ore is identified, mined, and the price and power follow the rock. Uranium is different. The fuel goes through four stages between the deposit and the reactor (extraction, conversion, enrichment, and fabrication), and the power to influence lies in the intermediate stages, in the hands of a limited bunch of countries and companies, not underground. This series sets aside the well-known simplistic formula of “more reactors, more mining” and traces the journey of the fuel throughout the entire chain, because the forces that truly drive uranium (both physical and geopolitical, and often hidden in the stages that most reports omit) are rarely found in the mine.

Kazakhstan produces about 40% of the world’s uranium,1 we probably all know that. It is repeated in every market note, every explainer, every nervous headline about supply security. It is, indeed, true and worth being aware of… But it answers the wrong question.

That 40% refers exclusively to mining, that is, extracting uranium from the ground as ore, which is milled into a coarse powder called yellowcake that is virtually useless as fuel without further processing. And almost nothing that makes the uranium market tight, strategic, or politically dangerous happens at the mine. It happens later, in a series of unglamorous industrial steps that transform raw uranium into something a reactor can actually use: Low-Enriched Uranium, or LEU, in which the concentration of the fissile isotope U-235 is raised to roughly 3–5% for the conventional reactors and, to just under 20% for the High-Assay Low-Enriched Uranium, HALEU, that the Chapter 2 of this series will cover. That middle of the chain is where the real concentration lies, and where, as we will see, Russia wields leverage. And it is where, over the past three years, a set of sanctions aimed at enrichment has raised global demand for mined uranium… without a single new reactor being built.

For these reasons, I have decided to explore this topic in depth and am therefore publishing this series, which I have decided to call the fuel-cycle squeeze series.

This first part covers the middle of the chain: enrichment, Russia’s grip on it, and a hidden dial (the tails assay) that silently sets how much mined uranium the world needs. The second chapter focuses on the fuel that the next-generation reactor fleet will use (HALEU) and on why next-generation reactors, artificial intelligence data centers, and space missions are about to compete for the same scarce supply. The third chapter returns to the topic of resources: how is it possible that the world has abundant underground uranium reserves and, at the same time, faces a supply shortage, and why is recycling presented as a solution when, in fact, it is currently not yet a solution. The fourth chapter discusses the mechanics from a market perspective: where each bottleneck is located, which companies are involved, and what factors would confirm or refute the hypothesis.

The four steps of uranium

Between a uranium mine and a working reactor are four industrial stages, and they are usually not interchangeable: each is a separate business, with separate plants, separate bottlenecks, and a very different geography:

• Mining produces yellowcake, the coarse powder of uranium oxide (U₃O₈) I mentioned in the introduction, which is the form uranium takes when it leaves the processing plant. This is the stage that most people tend to follow closely.
• Conversion transforms yellowcake into uranium hexafluoride (UF₆), a compound that sublimates (transitions from a solid to a gas) at moderate temperatures.
• Enrichment takes that UF₆ gas and raises the share of the fissile isotope, U-235, from its natural 0.7% toward the 3–5% a conventional reactor needs.
• Fabrication converts the enriched uranium into a solid ceramic (uranium dioxide, UO₂), presses it into pellets, and seals them into the metal tubes (fuel rods) that go into a reactor.

The geopolitical shape of these four steps is the whole story because it is wildly uneven.

Mining is indeed the most diffuse: It is spread across roughly a dozen countries, led by Kazakhstan (\~40%) and Canada (\~24%), and dominated at the company level by four producers: Kazatomprom (Kazakh state), Cameco (Canada), Orano (France), and CGN (China), which together account for more than half of world output. 1 But just look at who these four key players are: a Kazakh state-owned company, two Western companies, and one Chinese company, while the Russian mining subsidiaries (Rosatom’s Uranium One and ARMZ) are left just outside the group. No single bloc controls the mine on its own. Fabrication, at the far end, is similarly spread across about a dozen countries and currently sits in overcapacity, meaning the supply comfortably exceeds demand, prices are soft, and utilities can shop between suppliers 2, Uranium—Deposits, Production and Resources, Market Dynamics, and Supply Chain Risks. <em>Mineral Resources Program</em>. DOI: <a href="https://pubs.usgs.gov/publication/fs20253057/full" data-type="link" data-id="https://pubs.usgs.gov/publication/fs20253057/full">10.3133/fs20253057</a>. ") This is a beneficial market for buyers, not a true bottleneck, with one exception: Russian-designed VVER reactors require fuel assemblies specific to Russia, which historically have been manufactured only by Rosatom’s TVEL, a monopoly that Western manufacturers have only recently begun to break for Eastern European operators who inherited Soviet-era fleets. Between these two open ends are the two tight ones. Conversion runs through just five commercial plants on Earth: Canada (Cameco), China (CNNC), France (Orano), Russia (Rosatom), and the United States (ConverDyn) 2, Uranium—Deposits, Production and Resources, Market Dynamics, and Supply Chain Risks. <em>Mineral Resources Program</em>. DOI: <a href="https://pubs.usgs.gov/publication/fs20253057/full" data-type="link" data-id="https://pubs.usgs.gov/publication/fs20253057/full">10.3133/fs20253057</a>. ") three in Western-allied hands, one Russian, one Chinese. And enrichment, the most strategically sensitive step of all, runs through only four commercial suppliers worldwide: Russia’s Rosatom, the European consortium Urenco, France’s Orano, and China’s CNNC.3

Four suppliers, but one of them dominates the market, and this is Russia, which accounts for between 40% and 46% 4,5, Dependencies of the European Union and the world on Russian nuclear fuel cycle services, and how to reduce them. Energy Strategy Reviews. 62, 101923. DOI: <a href="https://www.sciencedirect.com/science/article/pii/S2211467X2500286X" data-type="link" data-id="https://www.sciencedirect.com/science/article/pii/S2211467X2500286X">10.1016/j.esr.2025.101923</a>."). That dominance stems mainly from the industrial infrastructure inherited from the Soviet era, which the West chose not to compete with. Rosatom operates four enrichment complexes with a combined capacity exceeding 27 million SWU per year,3 built during the Cold War and kept at full capacity throughout the 2000s by supplying the West with cheap enrichment services that made developing rival Western capacity seem economically unfeasible. The result is a supplier that is not only large but also structurally inexpensive: Russian enrichment has undercut the market for two decades, thereby undermining the incentive for any other player to expand. Replacing it, therefore, means not only matching its capacity but also rebuilding an industry that Russian competition had deliberately kept small, which is why, as we shall see, the transition takes years rather than months.

So the “40%” that should worry a Western utility isn’t Kazakhstan’s share of the mine. It’s Russia’s share of the step that turns uranium into fuel.

Why can’t you just dig up reactor fuel?

Natural uranium is a mixture of isotopes: atoms of the same element that are chemically identical but differ in mass because they have a different number of neutrons in their nuclei. Two of these isotopes are important when it comes to nuclear fuel. Uranium-235 is the fissile isotope: when it receives a neutron, it splits, releasing energy and more neutrons, which sustain a chain reaction. Uranium-238 is the inert ballast: in a conventional reactor, it basically just sits there doing nothing. The main issue is the enrichment level. Natural uranium contains approximately 99.3% U-238 and only about 0.7% U-235, and a conventional nuclear reactor requires the proportion of U-235 to be increased to approximately 3–5%. Therefore, before uranium can be used as fuel for anything, someone has to enrich the useful isotope: take a mass containing 0.7% U-235 and convert it into one containing 4 or 5%.

But, as I have already mentioned, U-235 and U-238 are the same element. They have the same number of protons, the same number of electrons, and the same chemical properties. It is not possible to separate them through any chemical reaction, since, from a chemical point of view, they are indistinguishable: there is no solvent, no membrane, and no reaction that favors one over the other. The only thing that differentiates them is their mass, and the difference in mass is just over one percent. Enrichment is, therefore, the industrial process of sorting atoms based on a one-percent difference in mass, repeated billions and billions of times.

The method that now dominates the industry is the gas centrifuge: that is, spinning uranium so fast that the heavier atoms edge outward from the lighter ones.

The reason why we need uranium in gas form and not just as a powder is that in its solid form, such as yellowcake (mainly U3O8\\mathrm{U\_3O\_8}​), uranium isotopes are locked into a rigid crystal structure. Under these conditions, physical processes like rotation cannot separate the isotopes, because the material behaves as a single mass rather than allowing individual atoms or molecules to move independently. For isotope separation to work, uranium must be in a phase where its particles can respond individually to external forces. This is achieved by converting uranium into uranium hexafluoride (UF6\\mathrm{UF\_6}​), a compound that becomes a gas at relatively moderate temperatures, and in the gaseous state, uranium exists as discrete UF6\\mathrm{UF\_6}​ molecules that move freely. When this gas is subjected to high-speed rotation in a centrifuge, a small but crucial mass difference comes into play: the molecules containing the heavier U-238 isotope tend to move slightly outward, while those containing the lighter U-235 concentrate closer to the center. This is why the conversion step exists: UF6\\mathrm{UF\_6} uniquely combines chemical stability with sufficient volatility, making it suitable for large-scale isotope separation using gas centrifuge technology.



#### Get the Raw Science Brief in your inbox.

Energy systems, critical materials, geopolitics..

The challenge is that the mass difference between U-235 and U-238 is extremely small, so a single centrifuge achieves only a minimal degree of separation. To reach useful enrichment levels, the process is repeated in interconnected sequences known as cascades. In these systems, the slightly enriched output from one centrifuge becomes the input for the next, with thousands of units operating in series and parallel configurations. Each stage incrementally increases the concentration of U-235, ultimately reaching the 3–5% enrichment required for most nuclear reactors. This is why enrichment is a four‑company business, not a four‑hundred‑company business. The centrifuges are slow, brutally precise, capital-intensive to build and qualify, and inherently dual‑use: the same machines that quietly turn out 3-5% reactor fuel can, with more stages and time, be driven up to roughly 90% weapons‑grade uranium. That reality puts the whole sector behind layers of nonproliferation walls (export controls, IAEA inspections, and treaty obligations) that decide who is allowed to own this capability. The supply shortage in this case is not a market coincidence; it is determined by the physical laws governing isotope separation and is further reinforced by nuclear safety policies.

SWU: the separation and enrichment unit

Enrichment is measured as effort, not as kilograms of uranium, so the industry uses its own unit: the separative work unit, or SWU, which does not tell you how much uranium you have but how much isotopic sorting you have done: conceptually closer to a unit of work than a unit of material. If you imagine natural uranium as flour full of mixed grains, enrichment is the sieve that separates the fine fraction (U‑235) from the coarse (U‑238). SWU measures how hard and how long the sieve is shaken, not how many kilograms of flour you started with or ended up with.

Every enrichment operation, then, produces three distinct material streams, and I would keep them clear since they are key to understanding the next sections:

• Feed: the natural uranium you put into the enrichment plant.
• Product: the enriched uranium you take out (the final 3–5% U‑235).
• Tails: the depleted uranium that comes out the other end, mostly U‑238 with a small residual amount of U‑235 that was not captured into the product.

The U‑235 fraction that you allow to remain in the tails, the so-called tails assay, is a genuine control mechanism. By choosing how lean or how rich you let the tails be, you change how much natural uranium the world must mine and convert to support a given amount of reactor fuel, even if reactor demand and operating conditions stay the same.

The dial: feed versus work

To make a given amount of reactor fuel, you can run the process two ways: You can run the tails lean: chase down nearly every atom of U-235 and leave very little behind. Do that, and you need less feed (you wasted almost nothing), but it costs you more separative work, because squeezing the last U-235 out of a stream that’s already mostly U-238 is exhausting work for the centrifuges. Or you can run the tails rich: you stop early, and leave more U-235 in the discard pile. Now you need more feed (you threw usable uranium away), but you spent less separative work to get there.

So feed and work trade off against each other. More work lets you get away with less uranium. Less work forces you to burn more uranium. The tails assay is simply where you choose to sit on that seesaw, and the rational place to sit depends entirely on which input is cheaper at the moment: mined uranium or enrichment capacity.

When enrichment is cheap and plentiful (as it has been in the era when large Russian centrifuge complexes could sell vast amounts of low‑cost SWU into the global market), the economics say: use lots of work, run the tails lean, and save uranium. Enrichers choose low-tails assays and spend extra separative work to strip more U‑235 out of each kilogram of feed, and in some cases even reprocess old tails piles to scavenge additional U‑235. Because contracts usually assume a higher transactional tails assay, an enricher operating this way consumes fewer tonnes of natural uranium than the paperwork implies and captures a surplus it can keep or sell.

This is known as underfeeding and behaves like a hidden secondary mine: it manufactures uranium‑equivalent out of extra work and effectively adds supply, pushing primary demand and prices down.

When enrichment is scarce and expensive (as in today’s Western market, where Urenco, Orano and others must replace sanctioned Russian SWU with more limited, higher‑cost capacity), the logic flips. Now separative work is the precious input, so you conserve it: you run the tails rich, accept more U‑235 going out in the waste stream, and feed in additional uranium to keep product volumes unchanged.

This is overfeeding and it creates additional demand for uranium that would not exist if enrichment capacity were inexpensive, which further exacerbates the shortage all the way back to the mines.

The reactors are structurally the same, and the fuel supplied is the same, but by turning this one dial (the actual tails assay used in the cascades), the world’s annual uranium requirement can move by many thousands of tonnes, almost entirely invisibly, through a setting on the enrichment plants that no headline ever reports.

Let´s see it in numbers

To give a sense of scale, Table 1 shows how much uranium feed and separative work are required to produce 1 kg of 4.5% enriched uranium from natural uranium (0.711% U‑235) at different tails assays. At a low tails assay of 0.15% U‑235, the plant would use about 7.75 kg of feed but must invest 8.28 SWU per kilogram of product; at a higher tails assay of 0.35%, it needs about 11.50 kg of feed but only 6.09 SWU.

🔒

Continue reading with Pro

This analysis is reserved for Raw Science Pro members.

• The physics of overfeeding: How a fractional shift in tails assays silently adds thousands of tonnes to global uranium demand without a single new reactor.
• The end of the Russian illusion: Why the secondary supply stack that kept uranium prices suppressed for a decade is permanently unwinding.
• The 2028 waiver cliff: The cascading effects of the U.S. import ban and why the Western enrichment gap is a highly leveraged demand event.
• The Chinese displacement loop: The observable trade data challenging the headline ban and its implications for the true scarcity of Western SWU.

References

1. World Nuclear Association, World Uranium Mining Production, 2026[](https://raw-science.org/uranium-enrichment-russian-ban/#easy-footnote-1-1993) 2. Mihalasky, M.J. USGS, (2025), Uranium—Deposits, Production and Resources, Market Dynamics, and Supply Chain Risks. Mineral Resources Program. DOI: 10.3133/fs20253057. [](https://raw-science.org/uranium-enrichment-russian-ban/#easy-footnote-2-1993) 3. World Nuclear Association, Uranium Enrichment, 2025.[](https://raw-science.org/uranium-enrichment-russian-ban/#easy-footnote-3-1993) 4. U.S. Department of Energy, Russian Uranium Ban Will Speed Development of U.S. Nuclear Fuel Supply Chain, 2024. [](https://raw-science.org/uranium-enrichment-russian-ban/#easy-footnote-4-1993) 5. Haneklaus, N. et al. (2025), Dependencies of the European Union and the world on Russian nuclear fuel cycle services, and how to reduce them. Energy Strategy Reviews. 62, 101923. DOI: 10.1016/j.esr.2025.101923.[](https://raw-science.org/uranium-enrichment-russian-ban/#easy-footnote-5-1993)