The Democratic Republic of Congo (DRC) produced more than 65% of the world's cobalt in 2023 reporting, and 15-20% of that Congolese tonnage still came from artisanal and small-scale mining (ASM) rather than industrial operators [S2]. That single geographic concentration is the starting point for every cobalt specification written in 2026, because price, ESG exposure, and supply continuity all flow from it.
Cobalt is not formally classified as a conflict mineral under the OECD 5-step framework, but it sits in the same risk envelope as tin, tantalum, tungsten and gold for downstream brands, and several majors, including Umicore, publicly exclude ASM cobalt from their feedstock [S2][S6]. The dominant end use is the lithium-ion battery cathode, with superalloys, cemented carbides, catalysts, and high-speed steels as the industrial fallback demand that keeps the market liquid when EV pull softens [S1].
What "cobalt raw material" actually means on a PO
Cobalt raw material entering a processor's gate is rarely pure Co metal; it is typically a hydroxide, sulfate, oxide, mixed Co-Cu concentrate, or recycled black mass, each with a different Co content window and impurity profile [S2][S7]. Benchmark Mineral Intelligence frames the supply chain in four commercial forms: mine concentrate, intermediate (hydroxide or sulfate), refined metal, and recycled feed, and spec writers should bind the form on the purchase order, not the element name [S7].
Umicore's 2020 disclosure illustrates the realistic mix a large refiner can disclose: roughly 75% of cobalt raw material from large-scale mining (LSM) in the DRC, the balance from other countries, and about 4.8% of total supplied cobalt from recycling [S2]. The recycled share is the variable to watch, because it changes the embodied-carbon figure that flows into a battery maker's cell-to-pack LCA.
Origin split: DRC dominance, mid-tier producers, and recycled feed
Around 50% of global cobalt reserves sit in the DRC, and Chinese and Indian mid-tier processors have been quietly adding DRC and Indonesian laterite-nickel (cobalt byproduct) capacity outside the headline Glencore Katanga flows [S5]. The mid-tier is the reason the cobalt market can move from shortage to surplus inside an 18-month window: Glencore's Katanga restart alone was sized to add roughly 30,000 tpa of cobalt output when it returned [S5].
For a 2026 spec, the practical origin map to lock into the tender is: (a) DRC LSM as the lowest-cost, highest-volume base, (b) Australian, Canadian, Cuban, and Philippine mid-tier mines as ESG-priced alternates, (c) Indonesian nickel-cobalt laterite as the fast-growing byproduct stream tied to NPI and MHP nickel projects, and (d) recycled cobalt from spent Li-ion batteries and catalysts as the ESG-premium source. Umicore's reported 4.8% recycled share in 2020 is a useful benchmark for what a "good" large refiner looks like before the EU Battery Regulation recycled-content thresholds start to bite [S2].
Due-diligence gates that the spec must enforce
Buyers who treat responsible sourcing as a CSR slide rather than a specification fail at the audit. Umicore operates a four-step due-diligence cycle, with mine-level traceability through lot receipts, chemical fingerprinting of impurity patterns on incoming lots, and alignment to the OECD Due Diligence Guidance for Responsible Supply Chains of Minerals from Conflict-Affected and High-Risk Areas [S2]. Jervois Finland's published cobalt policy layers the same OECD alignment onto nickel and other raw mineral materials, which is a reasonable template for a buyer's own supplier code [S8].
The Cobalt Institute's responsible-sourcing program and the Responsible Cobalt Initiative (RCI) led by CCCMC with OECD support are the two cross-industry schemes a procurement team should reference in tender pre-qualification [S3][S5]. Neither is a substitute for company-level due diligence; the Kumi analysis is explicit that industry initiatives are "welcome but not a substitute" for direct supplier audits, and that child-labor risk is only one of several material exposures, alongside worker safety at industrial processors, governance, and community relations [S5].
Specifications buyers should write into the contract
A defensible 2026 cobalt raw-material specification binds at least five parameters: (1) form and minimum Co assay by chemistry, e.g. Co(OH)2 at 30-40% Co, CoSO4·7H2O solution at 20-21% Co, or Co metal at 99.8% min; (2) impurity ceilings for Ni, Cu, Fe, Mn, Zn, Cd, and Pb, set against the cathode-chemistry the buyer is feeding; (3) moisture and loss-on-ignition limits for hydroxide shipments; (4) chain-of-custody documentation covering mine of origin, transporter, processor, and exporter, with lot numbers traceable end-to-end; and (5) third-party audit or chain-of-custody certificate, aligned to the OECD Due Diligence Guidance [S2][S7]. Umicore's "chemical fingerprinting" approach is a working example of how a buyer can verify origin claims against the actual impurity signature of a lot [S2].
For ESG and reporting, the spec should also require disclosure of the share of recycled input and the share of ASM, since the ASM percentage is the figure that maps directly to child-labor and human-rights exposure under the DRC context where ASM represents 15-20% of Congolese output [S2]. Buyers who cannot get an ASM/LSM split from a supplier should treat the absence of that data as a disqualifying finding, not a documentation gap.
Comparison: large-scale DRC vs mid-tier mines vs recycled feed
On four procurement-relevant criteria, the three main cobalt raw-material streams line up as follows. (1) Cost: large-scale DRC cobalt is the lowest unit-cost base, mid-tier Australian/Canadian material carries an ESG premium, and recycled feed is the most expensive per kg of contained Co but earns carbon and recycled-content credit value. (2) Volume stability: DRC LSM is the most volume-stable at the system level, but single-mine restarts (Katanga-class 30,000 tpa additions) can swing global balance fast, so the mid-tier and recycled streams matter precisely when LSM is disrupted [S5]. (3) ESG risk: DRC ASM carries the highest child-labor and human-rights exposure at 15-20% of Congolese output, mid-tier operations outside DRC carry lower ESG risk, and recycled feed carries the lowest social risk but higher chemical-handling risk [S2][S5]. (4) Lead time and form: mid-tier LSM ships hydroxide or concentrate with 6-12 week ocean lead times, recycled feed is regional and shorter-lead-time but variable in grade, and refined metal is the lowest-friction form for tight chemistry specs but the highest working-capital cost [S7].
Limits, failure modes, and what the spec cannot fix
No specification can fully hedge cobalt price volatility; Kumi documents that the cobalt market has historically moved sharply on supply restarts and on EV demand swings, with 2016 global demand around 96,000 tpa and a forecast doubling by 2025 that overshot on the way up and corrected on the way down [S5]. A buyer who writes a fixed-price annual contract for a security-critical application is accepting the risk that the spot price moves against them; the engineering-side countermeasure is to dual-source across at least two of the three streams (DRC LSM plus mid-tier, or mid-tier plus recycled).
Audit fatigue is the other structural limit. Industry initiatives such as the Responsible Cobalt Initiative, the Cobalt Institute's responsible-sourcing guidance, and the DRC Government's 2025 pledge to eliminate child labor in ASM are necessary scaffolding but they are explicitly not a substitute for direct supplier due diligence, and a single incident at a sub-tier processor can invalidate an otherwise clean chain of custody [S3][S5]. The Kumi analysis also flags dangerous working conditions at industrial (not artisanal) processors and exporters, which means an LSM-only sourcing claim is not automatically a low-risk claim [S5].
Adjacent raw-material context for battery buyers
Cobalt is almost never specified in isolation by a battery cell or cathode-active-material buyer, and the same DRC-and-Indonesia concentration logic now applies to lithium, nickel, and graphite procurement. For a parallel view on how spec-driven buyers in 2026 are gating lithium supply, see this lithium procurement strategy breakdown, and for a downstream look at how the upstream mix maps to Tier-1 cell makers, this Tier-1 cell supplier map is a useful cross-check. Where the cathode side touches the electric drive side, the e-axle magnet sourcing discussion in this e-axle procurement playbook shows the NdFeB and heavy-rare-earth side of the same risk picture, including the dysprosium and terbium exposure that cobalt-free LFP chemistry does not solve. [S5]
Cobalt is also a strategic material for non-battery uses, including superalloys for gas turbine engines, cemented carbides and diamond tools, and corrosion- and wear-resistant alloys, which is why the USGS maintains a dedicated Cobalt statistics and information page with monthly Mineral Industry Surveys, annual Mineral Commodity Summaries dating back to 1996, and the Cobalt Deposits in the United States data release [S1]. For more on the chemistry-side selection of battery raw materials, this battery-grade materials primer is a useful companion reference.
Trackable signals for the next buying cycle: the USGS Mineral Commodity Summaries 2026 cobalt chapter, due in early 2026, for the latest DRC share and global production tonnage; any update to the DRC Government's 2025 child-labor elimination pledge in ASM, originally logged in 2021 [S5]; and the share of recycled cobalt disclosed by major refiners in their next sustainability report, as the EU Battery Regulation recycled-content thresholds begin to push that figure well above Umicore's 2020 baseline of 4.8% of total supplied cobalt [S2].
Spec-level background on the components involved: linear guide, and crossed roller guide.