Cobalt Mining Supply Chain
and Ethical Concerns

DRC cobalt comes out of the ground on two tracks. Large-scale industrial mines, run by Glencore, China Molybdenum, Eurasian Resources Group and a few others, are mechanized and audited. The other track is artisanal and small-scale mining, ASM, where hundreds of thousands of people dig with hand tools in unsupported pits. No contracts. No safety gear. No legal standing. Amnesty International's 2016 report "This Is What We Die For" and subsequent IPIS field surveys put ASM at 15 to 30 percent of DRC cobalt output, a range that says something about how poorly the sector is tracked.

Cobalt prices surged from roughly twelve dollars a pound to over forty between 2017 and 2018, then crashed back below twelve in 2019. In both periods, child labor in ASM mining appears to have increased. Siddharth Kara, whose book "Cobalt Red" (2023) is based on fieldwork in Lualaba and Haut-Katanga between 2018 and 2021, documented this pattern across dozens of communities. During the price surge, the economic pull was overwhelming: a child's daily output was worth two to three dollars, which in southern DRC determines whether a family eats. During the crash, adult miners' earnings fell below household subsistence thresholds, and families responded by sending more members to the mines, children included, to aggregate enough small earnings to survive.

This is the part that most cobalt supply chain discussions get wrong, or rather, acknowledge for a sentence and then skip past on the way to talking about blockchain and due diligence frameworks. But the smelter is where the supply chain's ethical information system actually breaks down, and the breakdown is physical, not bureaucratic, and that distinction changes what solutions are even possible.

Start with the logistics. Cobalt ore leaves an ASM site and reaches a négociant, a local middleman, usually within hours. The négociant has ore from other sites. Everything gets mixed. That combined batch sells to the next intermediary, who mixes it with other batches. Two to five layers of this. By the time material reaches a smelter intake, the batch is a blend of unknown composition in terms of geographic and operational origin. No documentation connects any portion of it to a specific mine. The documentation, to the extent it exists, tracks weight and approximate grade. Not provenance.

At the smelter, cobalt concentrate goes into either a pyrometallurgical process (furnace temperatures above 1000°C) or a hydrometallurgical one (acid leaching, solvent extraction, precipitation). Both processes are destructive of the original mineral. Not destructive in the sense of damaging. Destructive in the sense that the input material's crystal structure ceases to exist and a new crystal structure forms. The output, cobalt sulfate or cobalt hydroxide, is a synthetic chemical product. It has a defined purity, a defined particle size distribution, a defined moisture content. It does not have a geographic origin encoded anywhere in its physical structure. There is no spectrographic test, no assay, no imaging technique available today that can determine, from a sample of cobalt sulfate, which mine the original ore came from.

This point tends to be acknowledged and then immediately abandoned in industry discussions, because dwelling on it leads to uncomfortable conclusions. If the information about extraction conditions is physically destroyed at the smelting stage, then everything downstream of the smelter is operating on trust. The OECD Due Diligence Guidance (2011, revised 2016) asks companies to identify and assess risk in the supply chain. The EU Battery Regulation (Regulation (EU) 2023/1542) requires supply chain due diligence. RMI's RMAP audits smelters. These mechanisms share a dependency on information flowing from mine to end user in some form. The smelter breaks the chain. After smelting, the only available information about feedstock origin is what the smelter volunteers about its own purchasing. The smelter is simultaneously the subject of external evaluation and the sole source of the data that the evaluation uses.

In other regulated sectors, this kind of structural conflict is mitigated through independent verification. Financial audits cross-reference company records against bank statements, tax filings, third-party transaction data. Food safety systems use physical testing at multiple points. Pharmaceutical supply chains use serialization and track-and-trace from API manufacturer to pharmacy. In the cobalt supply chain, there is no independent data source at the smelter level. The smelter says where its cobalt came from. The auditor checks what the smelter said. If the smelter is telling the truth, the system works. If the smelter is not telling the truth, there is no way to know from within the system.

Now, there is an alternative technical approach that could in principle bypass the smelter information wall, and it is worth discussing specifically because the industry has not invested in it proportionally to its potential.

Isotopic fingerprinting, piloted by Germany's Federal Institute for Geosciences and Natural Resources (BGR), exploits the fact that cobalt ore from different geological regions carries distinct trace element isotope ratio profiles. These profiles are products of geological processes spanning hundreds of millions of years. They are specific to regions, they are measurable in refined products, and they cannot be faked. A global reference database of these profiles, built through systematic geological sampling of cobalt-producing regions, would allow testing of any batch of refined cobalt product and inference of its geological origin. BGR's pilot work has established scientific validity.

Scaling this up has two practical obstacles. Building the reference database requires fieldwork across all major producing regions, and parts of the DRC, which is the most important producing region, are too insecure for systematic geological sampling campaigns. Mixed-origin smelter feedstock produces composite isotope signals that require high analytical resolution to decompose into constituent geographic contributions. These are engineering challenges. They are not conceptual challenges. The approach could work.

Around 2014, residents of the Kasulo neighborhood in Kolwezi found high-grade cobalt under their houses. The episode was documented by Bloomberg, Reuters, and in Amnesty International field work from 2015-2016. People dug through their own floors. Shafts went tens of meters deep without engineering. Houses collapsed. People died from cave-ins and gas accumulation in unventilated tunnels.

The significance beyond the immediate humanitarian catastrophe is geological. Mineralization in the DRC copper-cobalt belt does not respect concession boundaries. High-grade cobalt appears under residential neighborhoods, farmland, roads. The DRC's own mining cadastre (CAMI) has incomplete geological survey coverage, and where surveys exist, they often do not correspond to where ASM activity actually happens. IPIS monitoring pilots consistently report significant artisanal mining outside registered zones. A regulatory system designed around registered, geographically bounded mine sites captures a fraction of ASM activity. Possibly a small fraction.

Brief. The copper-cobalt belt overlaps with uranium mineralization. Shinkolobwe mine in Katanga supplied uranium for the Manhattan Project. Banza Lubaba Nkulu et al. (The Lancet, 2019) documented elevated urinary metal concentrations in mining community residents. ASM miners have no radiation awareness, monitoring, or protection. Health effects take ten to twenty years to manifest. Nobody with financial interests in the cobalt trade has an incentive to publicize this.

Audit arrival dates are known locally weeks in advance. Mines present compliance on audit day.

EGC, the Entreprise Générale du Cobalt, was created by the DRC government in 2019 as a monopoly buyer for ASM cobalt. EGC signed an exclusive offtake with Huayou Cobalt, one of the world's largest cobalt processors. Benjamin Faber and Benjamin Krause's 2023 NBER working paper on formalization interventions in Congolese mining found that past formalization efforts have had mixed results, with governance improvements frequently offset by elite capture. The DRC is near the bottom of the Transparency International Corruption Perceptions Index. Whether a state-created monopoly buyer in this governance environment constitutes labor rights enforcement or resource rent consolidation is a question that does not have a clear answer from available evidence. The structure is in place. What it produces for miners, as opposed to what it produces for the entities controlling the channel, is unknown.

The London Metal Exchange cobalt contract specifications are posted on the LME website. They specify minimum cobalt content for delivery: 99.3% for broken cathode, 99.8% for cut cathode. Origin is not a contract term. Whether the cobalt came from an audited industrial mine with safety officers and environmental management or from an unmonitored ASM pit where a twelve-year-old hauled ore in a sack does not affect the delivery price. If both meet the purity spec, they are fungible.

A DRC mine that invests in compliance absorbs those costs into its per-ton production number. Its cobalt does not sell for more. A mine that invests nothing in compliance has lower costs, higher margins, and sells into the same market at the same price. Compliance is a competitive disadvantage at the point of sale. The return on compliance investment materializes somewhere else entirely, at the consumer-brand level, where Apple or BMW convert responsible sourcing claims into brand equity. That return stays with the brand. It does not flow back upstream.

The Global Battery Alliance has discussed premium pricing. The Fair Cobalt Alliance has piloted above-market purchases. The LME has considered a separate delivery brand for responsibly sourced cobalt. None of these have reached meaningful scale. Trading houses and smelters that dominate physical flows resist supply chain bifurcation. An undifferentiated pool is operationally simpler. It persists.

The Faber and Krause NBER paper's finding that past formalization efforts in DRC mining have been undercut by elite capture makes sense in this context. Compliance has no market price. The only incentive to comply is external pressure. External pressure attenuates as it travels upstream through intermediary layers. At the mine face, it is background noise.

Topsoil stripping, erosion, heavy metal contamination of soil and water across ASM areas. Squadrone et al. (Environmental Research, 2020) measured elevated cobalt, copper, and uranium concentrations near the belt. Banza Lubaba Nkulu et al.'s Lancet work covers the health consequences. Acid mine drainage from industrial operations can contaminate groundwater for decades post-closure, sometimes much longer. DRC has relevant environmental legislation. Enforcement is negligible.

Cobalt concentrate from DRC ships to Chinese smelters, gets refined, goes to cathode plants in Korea or Japan, ends up in batteries sold in Europe or North America. The material crosses the globe roughly one and a half times. The shipping emissions do not usually appear in EV lifecycle assessments.

Chinese companies control roughly 70 percent of DRC industrial cobalt mining capacity and about 80 percent of global refining, per Benchmark Mineral Intelligence and the IEA. The U.S. IRA ties EV credits to sourcing rules penalizing Foreign Entities of Concern. The EU Critical Raw Materials Act (entered into force May 2024) targets domestic capacity building. Whether replacement refining capacity can materialize at scale within a decade is genuinely unclear.

Glencore shutting Mutanda in 2019, the world's largest cobalt mine at roughly 27,000 tons per year per Glencore's annual report, moved global supply more than the total cumulative effect of advocacy campaigns in the sector. Prices adjusted upward after the closure. Mutanda restarted in 2022. Oversupply returned. Prices fell. ASM miners took the brunt. Carter Center and Global Witness reporting on DRC mining revenue flows does not suggest royalty increases from the 2022 mining code revision are reaching mining communities in meaningful ways.

LFP batteries contain no cobalt and are now the highest-volume cathode chemistry in global EV installations. Sodium-ion batteries need no cobalt, nickel, or lithium and are entering early production.

If cobalt demand declines enough, ASM-dependent communities in southern DRC face income collapse. Formal employment in the region is scarce. DRC's social safety net does not function in any operational sense. A responsible cobalt-free transition strategy would include economic transition support for these communities. Major battery manufacturers' and automakers' ESG reports do not contain such a strategy.

The pricing structure on the LME is the load-bearing element. Ethical compliance has no market price. The governance architecture built on top of that pricing structure cannot produce outcomes that the pricing structure does not reward.