Abstract
Uranium recovery from wet-process phosphoric acid is the only mature industrial route to a critical raw material that was developed to commercial scale, operated for more than two decades, and then deliberately shut down on price. Between 1976 and 1999 plants in the United States reached a peak capacity of about 1,100 t U3O8 per year and accounted for roughly 90% of all commercial capacity of this type ever built worldwide. In some years the route supplied up to 20% of American uranium. It ended because the uranium price fell below the cost of recovery, not because the chemistry failed.
The conditions that justified that closure have all reversed. On 12 August 2026 the U3O8 spot price stood at USD 86.90 per pound, against demonstration-scale recovery costs from acid of USD 18-21 per pound. The United States consumes about 19,000 t of uranium a year and mined 534 t in 2025, roughly 3% of its own requirement, with about 97% imported; the statutory prohibition on Russian uranium has been in force since August 2024 and the waivers that still carry part of the supply expire on 1 January 2028. World mine production was 60,213 t U in 2024 and covered only 90% of reactor demand, with 39% of it from a single country.
This paper argues that the constraint on this route has never been geological and is no longer economic. It is the form of the plant: conventional solvent-extraction and ion-exchange trains are large, site-specific capital projects with twenty-year payback horizons, and that horizon has already disappointed investors once. We describe ARBOK-SA – cold boiling under deep vacuum at ambient temperature, with separation in a single pass inside the solution itself – as a containerised alternative that turns a capital project into equipment, and we quantify the resource at site, national and global scale.
1. Where the uranium actually is
Uranium in phosphate rock is a companion element. It substitutes into the apatite lattice and is distributed through the ore at 50-200 mg/kg, with individual deposits above 400 mg/kg. It cannot be concentrated by beneficiation or flotation: the grade is far too low for the rock to be treated as a uranium ore, and too high to be dismissed.
The decisive fact appears in the wet process. Phosphate rock is digested with sulphuric acid; calcium sulphate precipitates as phosphogypsum, about 5 t per tonne of P2O5. Uranium does not follow the solid. Approximately 80-90% of the uranium in the ore reports to the liquid phase – to the wet-process phosphoric acid itself – and only 10-20% remains in the gypsum.
This inverts the usual problem. The uranium does not have to be found, no rock has to be broken for it, and no mine has to be permitted. It is already dissolved, already in an acid medium, and already being pumped through the pipework of an operating plant. From a recovery standpoint this is close to an ideal feed: a homogeneous solution of stable composition, a steady flow, and continuous operation.
There is a second consequence. Uranium that stays in the acid does not disappear. It follows the process into monoammonium and diammonium phosphate and into superphosphate, and from there onto agricultural land. No limit value for uranium in fertiliser exists today in either the European Union or the United States. The question surfaces periodically and remains open. For a phosphate producer this means that uranium in the product is neither an asset nor a liability, but a neutral quantity nobody asks about – until somebody does.
2. What was built, and why it was stopped
The United States understood this earlier than anyone and built the plants. Recovery units operated from 1976 to 1999 with a peak aggregate capacity near 1,100 t U3O8 per year. Historical operating costs were reported at USD 11-45 per pound U3O8 in 1979 dollars.
Then everything closed. The reason was singular and had nothing to do with the technology: uranium became cheap. With the pound trading near USD 10, no recovery cost in that range could survive. Plants were dismantled, staff dispersed, and the operating knowledge decayed. An industry supplying a fifth of the country’s uranium ceased to exist not because it did not work, but because it was more expensive than imports.
This is worth stating plainly, because it is unusual. Uranium from phosphoric acid is the only large industrial recovery route for a critical material that was carried to maturity, run for decades, and then deliberately retired on price. It did not fail. It was switched off.
3. What changed by 2026
Price. The U3O8 spot price on 12 August 2026 was USD 86.90 per pound. Demonstration-scale estimates of operating cost for recovery from acid are USD 18-21 per pound. The gap is fourfold and has not been a momentary excursion.
Access. The United States consumes about 19,000 t U per year and produced 534 t from its mines in 2025, up from 260 t in 2024 and from lows of 6-88 t in 2020-2022. Domestic output covers roughly 3% of requirement. The prohibition on imports of Russian uranium products has applied since August 2024 and runs to the end of 2040; waivers permitting limited low-enriched uranium imports expire on 1 January 2028.
Concentration. World mine production in 2024 was 60,213 t U, of which Kazakhstan supplied 39%, Canada 24% and Namibia 12% – three countries for three quarters of the market. Mine output covered 90% of reactor requirements, the balance coming from inventories and secondary supply. More than half of world uranium is now produced by in situ leaching, which means the industry already works with solutions rather than rock. The premise that uranium implies a mine is obsolete.
4. Why the old route has not returned
The classical answer requires a plant inside a plant. Solvent-extraction or ion-exchange recovery from phosphoric acid means a large capital installation with its own reagent inventory, organic-phase circulation, regeneration cycles and dedicated operating staff. It cannot be installed on one site and moved to another. It is built for a specific acid plant on a twenty-year view – and that view has already failed once, within living memory of the industry it would serve.
The result is a standoff. The feed runs through the pipe, the price is four times the cost of recovery, and capital does not enter the niche because it remembers 1999.
5. ARBOK-SA
ARBOK-SA works with the feed in the state it is already in. The stream is subjected to cold boiling under deep vacuum at ambient temperature. There are no furnaces, no membranes and no purchased reagents in the core of the process. Separation occurs in a single pass, inside the solution itself, without the classical refining cascade of hundreds of stages. Different physics wins not on volume but on the singularity of the approach.
The practical consequence for a phosphate plant is not that a kilogram of uranium is cheaper than on a large sorption train. It is that the unit arrives containerised, is placed on an existing stream, and requires neither the construction of a process building nor a twenty-year payback horizon. Modularity converts a project of the century into equipment.
The same equipment is not limited to acid. Uranium exists in liquid form in four streams, and every one of them is already somebody’s operating cost:
- wet-process phosphoric acid, carrying 80-90% of the uranium in the ore;
- phosphogypsum, fresh and stacked, carrying the remaining 10-20%, against a liability that is already funded;
- mine and quarry waters of uranium and coal basins, treated under reclamation obligations;
- in situ leach solutions, by definition a liquid phase and already more than half of world uranium output.
6. Who the buyer is
The customer for such a unit is not a uranium company. It is a phosphate fertiliser producer through whose plant uranium currently passes in transit, earning nothing. No change to the core process is required, no rebuild of the sulphuric acid or ammonia section, and no new market to learn: uranium concentrate is an exchange-traded commodity with established contracting and a queue of buyers. This is a second product from the same feed, on the same site, with the same staff.
7. Economics
Consider a site handling 1 Mt of phosphogypsum a year. That corresponds to 200,000 t P2O5, about 667,000 t of rock, and roughly 510,000 m3 of wet-process phosphoric acid a year – 1,400 m3 per day, or 7 standard ARBOK units of 200 m3/day.
The uranium in that stream is approximately 31 t a year. At current pricing this is on the order of USD 7 million a year of contained value from a single site, against recovery costs of roughly USD 1.6 million at demonstration-scale unit costs.
This is the first case in the ARBOK critical-materials series in which the target element pays for the project by itself. Terbium, lanthanum, lutetium and erbium emerge from phosphogypsum as a basket, and none of them individually funds an installation. Uranium does, and it makes everything else recovered from the same pass an addition rather than a requirement.
Nationally the picture is sharper. Against United States phosphate rock production of 20 Mt a year, on the order of 935 t of uranium reports to the acid – approximately USD 210 million a year at current pricing, and nearly twice the country’s entire operating mine output. It does not replace imports: 935 t against a 19,000 t requirement is about 5% of demand. It does double domestic production without a single new mine, exploration programme, land withdrawal or pit.
At world scale, roughly 13,900 t of uranium a year passes through phosphate processing; some 9,000 t of that sits in streams above a 90 mg/kg cut-off, which is a working range rather than a theoretical one. Nine thousand tonnes is about 15% of world mine production and more than Namibia, the third largest producer on the planet, supplies from ore.
8. Limits of the claim
Uranium is radioactive feedstock and its handling is licensed. That is separate work for specialist operators and is not undertaken here. ARBOK equipment delivers a uranium fraction from the stream; it does not deliver commercial-grade yellowcake and it certainly does not deliver fuel. Upgrading the fraction to a saleable specification, and everything downstream of that, is another party’s process step and is not costed in the figures above.
Uranium content in phosphate rock varies by a factor of four between deposits. A site cannot be evaluated against an average; direct determination of uranium in the specific stream is a precondition.
Finally, the price of uranium has already destroyed this industry once. Project economics must be built on long-term contracts rather than on spot, or 1999 repeats.
9. Conclusion
Florida holds 25 phosphogypsum stacks containing roughly 1 billion tonnes of material, with about 30 million tonnes added every year. Through the same sites runs phosphoric acid carrying the dissolved uranium of a country that imports 97% of its uranium. The acid is a product: it is already pumped, cooled, filtered and moved onward.
The question is not geological and it is not about the price of the metal. The question is that between the dissolved uranium and the market stands one piece of equipment which is not present on those sites.
Forty years ago it was there. It was removed because it stopped paying. Today the thing that does not pay is leaving it out.
ARBOK Strategic Research Institute
CC BY 4.0