The 10,700-Ton Hole in Glencore’s Cobalt Supply

A cathode materials procurement manager in Shenzhen stopped waiting for Glencore’s containers from Kolwezi in February. Now she places weekly cobalt orders from a warehouse 120 kilometers away, on the Wuxi Stainless Steel Exchange, competing with every other buyer whose African supply just got rationed.

22,800 Tons and a 10,700-Ton Hole

The Democratic Republic of Congo produces roughly 72% of the world’s cobalt, according to Our World in Data — the silvery-blue metal that stabilizes cathodes in most EV batteries. In early 2026, the DRC shifted from a blanket export suspension to a quota system. As reported by Ecofin Agency, Glencore, the world’s largest cobalt trader, produced 33,500 tons from its Congolese operations the previous year. Under the new quotas, it can export 22,800 tons. That leaves a 10,700-ton shortfall — roughly 4% of global supply — that Glencore must fill from somewhere else.

Where it found the cobalt is the story.

DRC’s shift from export suspension to quota-based rationing, combined with assay disputes and mine collapses, has forced the world’s largest cobalt trader to source from Chinese exchange stockpiles—revealing that Western EV battery supply chains now depend on Chinese inventory access for critical mineral fulfillment.

Glencore turned to stockpiles on China’s Wuxi Stainless Steel Exchange, a commodities platform most Western investors have never heard of. Wuxi cobalt inventory has fallen to 3,934 tons, down from more than double that figure in late January. The drawdown suggests Glencore is not the only trader pulling from the same shrinking pool.

Western mining companies invested billions in DRC operations specifically to secure cobalt outside Chinese control. Glencore operates its own mines in Kolwezi. And now it fulfills Western battery contracts by purchasing metal from Chinese exchange warehouses — because the DRC’s quota system made direct exports insufficient. The quota restrictions redirected even a vertically integrated miner to the Chinese trading infrastructure it had sought to avoid.

Can Glencore simply wait for quotas to loosen? Its own disclosure suggests not — excess cobalt “continues to be stored in-country and will be sold as circumstances allow,” meaning physical metal sits in DRC warehouses with no export pathway while contracts come due in Stuttgart and Detroit.

The M23 rebel group’s control of the Rubaya coltan mines since 2024 has created a parallel extraction economy operating entirely outside Kinshasa’s quota enforcement. Ibrahim Taluseke, a miner at the Rubaya site, told Al Jazeera he helped recover over 200 bodies after a March landslide, adding that “the owners of the pits do not accept that the exact number of deaths be revealed.” That opacity extends to export volumes — material leaving rebel-controlled territory does not appear in official quota tallies. If significant cobalt volumes leak through these informal channels, the effective supply reduction may be substantially smaller than the 10,700-ton shortfall implies. No reliable data exists to quantify these flows.

The quota didn’t reduce cobalt’s availability. It relocated who controls access.

Trafigura Bets $1.1 Billion on Battery Graveyards

That relocation of control explains why a different commodity trader made a move that would have seemed premature two years ago. In March, Trafigura signed a binding $1.1 billion offtake agreement with Nth Cycle, a recycling company based in Fairfield, Ohio. The deal commits Trafigura to purchase 3,500 tonnes annually of recycled nickel and lithium extracted from black mass — the shredded residue of spent EV batteries.

Trafigura’s binding $1.1B offtake agreement with Nth Cycle for 3,500 tonnes annually of recycled nickel and lithium from battery waste signals that major commodity traders now view closed-loop recycling as commercially viable at scale, potentially weakening the pricing leverage of primary mining nations like Indonesia and the DRC.

Black mass contains cobalt, nickel, lithium, and manganese in a mixed powder that, until recently, was too expensive to refine competitively. Nth Cycle’s Oyster process uses electrochemical separation rather than smelting or acid leaching, claiming up to 70% lower capital intensity.

The skeptic’s objection is obvious: 3,500 tonnes is a rounding error against global demand. The deal doesn’t deliver material until 2028, requiring 12,000 tonnes of black mass feedstock annually — a collection infrastructure that barely exists. But the signal is not the volume. It is the counterparty.

Trafigura is not a cleantech venture fund. Tomoya Murata, Trafigura’s general manager in Japan, signed the agreement in Tokyo during the Indo-Pacific Energy Security Ministerial, with U.S. Secretary of the Interior Doug Burgum in attendance. When a trader whose business model depends on primary extraction locks in a billion-dollar recycling contract at a diplomatic forum, the commercial logic has already shifted.

The Confession in the Contract

For decades, commodity traders built their edge on privileged access to primary extraction: mine-gate offtakes, port logistics, sovereign relationships. The entire business model assumes that controlling upstream supply is the durable advantage.

Trafigura’s $1.1 billion commitment to recycled feedstock demonstrates that a major trader now views secondary material supply chains as commercially viable at scale. That is a shift from treating recycling as marginal.

According to Investing News Network’s year-end review, cobalt metal prices rose from near nine-year lows of $24,343 per metric ton in January 2025 to $53,005 by December — a 118% increase. Glencore’s pivot to Wuxi stockpiles showed that even the largest miner-trader in the world cannot guarantee physical delivery from its own operations. If Glencore — with its own mines, its own rail, its own ships — ends up buying from Chinese warehouses, then every smaller battery maker faces the same dependency with less leverage.

The DRC quota crisis did not cause this belief. It confirmed it.

The binding constraint moved rapidly in early 2026 — from DRC mine output to DRC export policy, then from export policy to Chinese exchange inventory. Each shift happened because the actor holding the constraint adapted rationally: the DRC imposed quotas to maximize revenue per ton; Glencore tapped Wuxi to honor contracts; Chinese exchange operators accumulated inventory before restrictions tightened. No single actor caused the dependency. Each one’s rational adaptation deepened it.

Trafigura’s recycling offtake is the first bypass — a route that, if it scales, eliminates the Chinese inventory chokepoint entirely by sourcing battery metals from spent batteries rather than from mines or exchanges. But the bypass reveals the next binding constraint: electrochemical processing infrastructure. Nth Cycle’s Oyster technology requires 12,000 tonnes of annual black mass feedstock, but the collection networks, sorting facilities, and electrochemical separation plants don’t yet exist at commercial scale in the United States or Europe. The constraint migrated from feedstock access — which the DRC controlled — to deployment infrastructure for electrochemical recycling systems, which nobody has built yet.

The procurement manager in Shenzhen still places her weekly Wuxi orders. The constraint moved from geological deposits to processing capacity, but the infrastructure to exploit that shift hasn’t been deployed.

What to Watch

Wuxi Stainless Steel Exchange cobalt inventory reports, published monthly, are the clearest pressure gauge. If stocks fall below 2,000 tonnes by July 2026, Glencore and other traders face a physical delivery crisis that no financial instrument can paper over — though rising prices will likely trigger demand destruction or substitution before inventories hit zero.

I predict DRC government statements on quota adjustments will be the next catalyst. Any relaxation before Q3 2026 signals that Kinshasa’s revenue losses from restricted volumes have exceeded the pricing gains from artificial scarcity. If Kinshasa holds firm through the summer, the government believes it can outlast the traders.

On the recycling side, Nth Cycle’s Fairfield plant commissioning timeline matters more than any policy announcement. If the company breaks ground on its second facility by October 2026, it confirms that Trafigura’s offtake economics work at the unit level. Watch also for competing black mass offtake announcements from Glencore or CMOC — a second major trader entering recycling would confirm the pattern is structural.

If This Thesis Is Wrong

First, the LFP substitution trend may accelerate faster than anyone projects. If cobalt-free chemistries capture 70% or more of new EV production by 2028, the entire cobalt supply chain becomes a sideshow. The Trafigura deal would then look like a hedge on a declining market, not a signal of structural change.

Second, the DRC’s quota system may prove temporary. If the policy costs the government more in lost volume than it gains in pricing power, a reversal could come within months. A quota relaxation by mid-2026 would refill Glencore’s export pipeline and deflate the Wuxi drawdown story entirely.

Third, Glencore’s Wuxi purchases may reflect routine inventory management rather than desperation. Large commodity traders constantly arbitrage between exchanges. The strongest version of this counterargument: Glencore has operated in volatile African jurisdictions for decades and has always found workarounds. Why should this time be different?

Every satisfying supply chain story has a gate that moved and a winner who saw it first. This one has a constraint that migrated from geological access to electrochemical infrastructure — and nobody has deployed the processing capacity to exploit it yet.

The metal comes from Congo, but the gate is in Wuxi.

Market Implication

If this thesis holds, LME cobalt futures express the Wuxi inventory drawdown directly—physical delivery pressure builds as stocks approach 2,000 tonnes by July. A prediction market framing: “Will Wuxi cobalt inventory fall below 2,000 tonnes before August 2026?” captures the article’s core inflection point. The sleeper trade is exposure to electrochemical separation equipment manufacturers supplying Nth Cycle’s technology—Trafigura’s $1.1B offtake proved recycling economics work, but the processing infrastructure to exploit that shift hasn’t been deployed yet. Kill signal: DRC quota relaxation before Q3 2026 or Chinese strategic reserve releases refilling exchange stocks.

Analytical implication, not financial advice.

Sources