Skip to content

Science1 publisher3 min readPublished

Geologic modelling puts 2,000 to 5,000 tonnes of uranium in DRC's cobalt exports since 2000

A study of Copperbelt geology and plant throughput estimates that 25 years of cobalt hydroxide shipments carried natural uranium, less than a tenth of it declared to the IAEA. The DRC is now testing outgoing loads.

The Scientist · Science desk

Photograph accompanying Geologic modelling puts 2,000 to 5,000 tonnes of uranium in DRC's cobalt exports since 2000
Photo: nature.com

What happened

  • A study of Copperbelt geology and processing estimates that 2,000 to 5,000 tonnes of natural uranium left the DRC inside cobalt hydroxide shipments between 2000 and 2024 in the most likely scenarios.
  • Less than 10% of that material appears in public declarations to the International Atomic Energy Agency, with most of the shipments going to China, which dominates the global cobalt market.
  • In the weeks after the study appeared in July 2026, the DRC launched a probe to test outgoing cobalt shipments for uranium and said it would consult the IAEA.

Compiled by The ScientistSomething wrong?How this is made

Why it matters

  • constraint With a high end 2.5 times the low end and no shipment assays behind it, the figure can size a national gap over 25 years but cannot support a claim about any single cargo or plant.
  • exposure Refiners buying DRC cobalt hydroxide are the ones holding whatever uranium travels with it, and the chemistry for stripping it has only been discussed since about 2020 and is not in common use.
  • decision Safeguards accounting starts from declared quantities, so a declaration gap this size forces the IAEA and the DRC to decide whether cobalt hydroxide belongs in the reporting system at all.
  • capability Testing outgoing shipments would turn a modelled tonnage into a measured one, and give buyers a number they can ask a supplier for.

The study's two headline conversions are rescalings of a single number. Divide the low tonnage by the low weapons count, 2,000 over 600, and you get about 3.3 tonnes of natural uranium per weapon; do it at the top of the range, 5,000 over 1,500, and you get 3.3 again [3]. The reactor figure works the same way, with 2,000 tonnes over 10 years and 5,000 over 25 both coming to 200 tonnes a year [4]. So the 600 to 1,500 weapons range carries no information the tonnage did not already carry, and it says nothing about who could enrich the material or whether anyone tried [4][5].

Spread across 2000 to 2024, the estimate averages 80 to 200 tonnes of natural uranium a year [1]. Less than 10% of the total appears in public declarations to the International Atomic Energy Agency, the body responsible for uranium accountancy and safeguards [2][18]. That leaves 1,800 to 4,500 tonnes outside them [2]. The DRC has not officially reported any uranium exports in decades [15].

The number comes out of a model. The authors used geologic maps to estimate likely uranium grades in cobalt ores, combined those grades with the ore throughput of each Copperbelt facility, and added a model of how uranium behaves in processing chemistry before shipment [6]. They did that because very few data points on uranium concentrations in the ores, or on the chemical steps used to prepare cobalt for export, are public [7]. A calculation of this shape can bound a national total over 25 years, though it cannot tell an importer what is in one container. The high end of the range is 2.5 times the low end [5].

The mechanism is the part with independent support. The region's concentrated uranium deposits formed roughly 600 million years ago, and over the past few million years rainwater has carried uranium out of them across the Copperbelt [10]. Heterogenite, the mineral that supplies most of the cobalt, can adsorb that uranium as it passes [11]. In the crude refining done inside the DRC, uranium and cobalt behave alike, so nearly all the uranium that comes out of the ground stays in the export product [12]. Refineries can strip it out, but discussion of those techniques began only around 2020 and they are not common [13]. The authors also compiled DRC import records for the chemicals those processes require [14].

"The specific risk of uranium exports as a byproduct of cobalt is not a story of covert access to deposits and smuggling. It stems from a geologic coincidence," the study's author, a geologist writing in Phys.org, wrote [19]. Southern DRC supplies 70% of the world's cobalt [9]. Shinkolobwe, the Manhattan Project's main uranium source, has been officially closed to excavation since 2004 [17].

What the source does not document is any testing regime among buyers: there is no survey here of what importers assay on arrival. What it does document is the probe the DRC launched in the weeks after the July 2026 publication, which will test outgoing cobalt shipments for uranium content and consult the IAEA [16]. Those assays would supply the one quantity the model estimates instead of observing, the uranium concentration in material actually leaving the country.

What to watch

  • Results from the DRC's shipment testing, and whether measured uranium concentrations fall inside the modelled 2,000-5,000 tonne band.
  • Whether the DRC begins reporting uranium exports to the IAEA, and whether past cobalt hydroxide tonnage is reconciled retroactively.
  • Whether refiners taking DRC cobalt hydroxide add uranium removal steps, which would show up in import records for the required chemicals.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories