A June 2026 study in Environmental Science and Ecotechnology, mapping cobalt flows across 230 countries and six supply-chain stages from 1998 to 2019, finds that the network of potential failures is four times denser than the visible physical trade system. Risk concentrates at mining but accumulates most dangerously at refining and manufacturing bottlenecks, producing nonlinear, abrupt collapses rather than gradual declines. The findings arrive as DRC export quotas have already halved allowable cobalt volumes, cobalt metal prices have risen more than 160% from their February 2025 lows, and Western supply-chain strategies remain structurally misaligned with the systemic nature of the threat.
Introduction
A peer-reviewed study published in Environmental Science and Ecotechnology and highlighted by ScienceDaily on June 20, 2026 has delivered one of the most methodologically rigorous quantifications yet of systemic risk in the global cobalt supply chain. The research, authored by Xin Ouyang, Litao Liu, Qiance Liu, Wu Chen, Chao Wang, Xun Pang, Canfei He, and Gang Liu from institutions including the Chinese Academy of Sciences, Peking University, and the University of Southern Denmark, mapped cobalt flows across 230 countries and six interlinked supply-chain stages from 1998 to 2019. Using a multilayer shock propagation model integrated with trade-based material flow analysis, the authors produced the most granular picture to date of how a single disruption propagates through the global cobalt network.
The headline finding is stark. The resulting avalanche network of potential failures is roughly four times denser than the physical trade network itself, meaning that a targeted shock at a critical node does not produce a proportionate, linear response. It produces a cascade. Disruptions travel through alternating direct and indirect pathways, moving horizontally across international trade links and vertically across production stages simultaneously, prolonging interruptions and generating nonlinear ruptures that traditional country-level risk assessments are structurally unable to detect.
The study's timing could not be more consequential. As of mid-2026, the Democratic Republic of the Congo has replaced its 2025 export ban with a quota system capping annual cobalt exports at approximately 96,600 tonnes, roughly half of 2024 production levels. Cobalt metal prices have risen more than 160% from their February 2025 lows to above US$56,000 per metric tonne at the start of 2026, and cobalt hydroxide payables surged approximately 350% between February 2025 and April 2026. The paper's central argument that the global cobalt supply chain is robust to random small shocks but catastrophically fragile to targeted disruptions at high-concentration nodes is not a theoretical concern. It is the lived reality of current market conditions.
Methodology and Core Findings: The Avalanche Network Defined
The study's methodological contribution rests on its integration of two analytical frameworks that are typically applied separately. Trade-based material flow analysis maps physical cobalt volumes across the six life-cycle stages (mining, refining, manufacturing, use, waste management, and recycling) and across 230 countries and regions. The multilayer shock propagation model then simulates how supply shortages or demand collapses at any single point in that network spread through the broader system, both within production stages and across them.
Systemic risk is quantified using two indicators: systemic fragility, which measures a node's capacity to trigger widespread collapses if disrupted, and exposure rate, which measures a node's vulnerability to collapses originating elsewhere. The combination of these two indicators allows the authors to distinguish between countries that are systemically dangerous (high fragility, meaning disruptions there cascade outward) and countries that are systemically vulnerable (low fragility but high exposure, meaning they absorb shocks from elsewhere without the capacity to absorb or redirect them).
The core structural finding is the characterisation of the cobalt supply network as robust-yet-fragile. The network can absorb numerous small, distributed, random shocks without systemic failure. This robustness, however, is deceptive. When disruption targets the critical nodes where concentration is highest, the network collapses rapidly and nonlinearly. The avalanche network of potential cascading failures is approximately four times denser than the visible physical trade network, revealing extensive hidden interdependencies that are invisible to analysts examining only direct trade relationships.
The directional finding on where risk concentrates is particularly important for policy. Mining disruptions, particularly in highly concentrated upstream regions, are frequent risk sources. However, the most severe systemic impacts accumulate at refining and manufacturing bridges, where dense vertical and horizontal connections amplify failures. A disruption at the DRC mining stage is the most common shock scenario, but it is at the refining and manufacturing layer that the failure energy concentrates and propagates at maximum force.
Supply Concentration Reality: Why the Model's Findings Are Not Abstract
The study's abstract architecture maps onto a supply landscape of extraordinary physical concentration. According to USGS data, the DRC produced approximately 230,000 tonnes of mined cobalt in 2025, representing roughly 73% of the global total of 310,000 tonnes. Indonesia was the second-largest producer at approximately 14%. The upstream mining stage, which the study identifies as the most frequent risk source, is therefore dominated by a single country to a degree that has few parallels among critical minerals.
The refining and manufacturing layer, which the study identifies as the stage where cascading failures accumulate most severely, is dominated to a comparable degree by a single country at the processing stage. China accounted for approximately 78% of global refined cobalt production in 2024, according to IEA data. Roughly three-quarters of global cobalt sulphate and metal output passes through Chinese facilities regardless of where the ore originates. Chinese companies own or finance approximately 15 of the 17 largest cobalt mining operations in the DRC, meaning that Chinese capital controls the upstream feedstock and Chinese industrial infrastructure controls the downstream transformation.
This vertical integration from DRC geology to Chinese refining capacity is precisely the structural configuration that the study's model predicts will generate maximum cascade severity when disrupted. The refining and manufacturing bridges are not distributed across diverse geographies; they are concentrated in a single economy that simultaneously represents one of the two nodes the authors identify as exhibiting high systemic fragility. China and the United States are both flagged as countries whose disruptions can trigger widespread collapses given their centrality to the network. The asymmetry is that China's fragility is coupled with control, while the United States' fragility reflects dependency.
A further structural complication is cobalt's byproduct status. Approximately 94% of cobalt is produced as a byproduct of copper (roughly 50%) and nickel (roughly 44%) mining, with only about 6% from dedicated cobalt operations. This means cobalt mine production responds primarily to copper and nickel market economics rather than to direct EV battery demand signals. A collapse in copper or nickel prices can restrict cobalt output even when battery demand for cobalt is growing, adding a layer of market dislocation that the study's shock propagation model is specifically designed to capture.
The DRC Quota System as a Live Stress Test of Cascade Dynamics
The DRC's policy trajectory since February 2025 provides a real-world experiment in precisely the kind of targeted upstream shock the study models. The initial export ban, implemented in February 2025 and replaced in October 2025 with a structured quota system, removed a substantial volume from global supply. The 2026 annual quota of 96,600 tonnes, of which 87,000 tonnes is distributed to producers on a pro rata basis with 9,600 tonnes retained under the state regulator ARECOMS' discretionary control, represents approximately half of the DRC's 2024 production volume of an estimated 230,000 tonnes.
The cascade effects have been visible and measurable. Cobalt metal prices entered 2026 above US$56,000 per metric tonne, more than doubling from their 2024 lows. Cobalt hydroxide payables rose approximately 350% between February 2025 and April 2026, from US$5.60 to 5.75 per pound to US$25.90 to 26.00 per pound. Fastmarkets senior analyst Robert Searle noted that the quotas represented a significant cut on pre-ban monthly volumes and forecast deficits of around 5,000 to 6,000 tonnes for both 2026 and 2027 as Chinese refineries contend with tighter intermediate supply. Benchmark Mineral Intelligence's Roman Aubry went further, warning of significant upside price risk as dwindling ex-DRC stocks present the risk of demand destruction toward the end of 2026.
Logistical failures have compounded the policy shock in ways that align with the study's cascade propagation mechanics. A source familiar with DRC export border documentation indicated that from December 2025 through the end of February 2026, only 7,800 tonnes of cobalt had been cleared for export, against a quota that implies substantially higher throughput. A DRC-based logistics source estimated that less than 50% of the Q4 2025 and Q1 2026 quota had been filled, citing trucking disruptions including a bridge collapse on a key export route and paperwork compliance requirements under the new 10% advance royalty rule. Shipping cobalt hydroxide from the DRC to Chinese refineries typically requires three to four months, meaning these logistical failures will transmit delayed but concentrated tightness into refining feedstock availability later in 2026, exactly the kind of amplification at the refining bridge that the study's model predicts.
The quota's internal design flaws add a further layer of instability. Elisabeth Caesens, founder of Resource Matters and speaking at the Cobalt Institute's 2026 annual congress in Madrid, identified a central contradiction: the quota methodology rewards historical production volumes, meaning the companies whose aggressive output growth between 2022 and 2024 contributed most directly to the oversupply problem now receive proportionally larger allocations. The formula runs counter to the policy's own stated stabilisation goals. MMG's Kinsevere operation received a 2026 quota of just 360 tonnes against a facility capable of producing up to 6,000 tonnes annually, a 94% shortfall that General Manager Aaron Chen characterised as rendering cobalt production economically unviable. Panmure Liberum analyst Duncan Hay summarised the structural uncertainty in market terms: Congo's shifting export rules offer no certainty, and last-minute royalty demands and complex paperwork will keep exports and prices volatile.
Policy Implications: Why Stage-Aware, Multilateral Coordination Is Structurally Necessary
The study's policy conclusions are direct and unambiguous. Traditional country-level risk assessments underestimate the true vulnerability of the cobalt supply system because they evaluate countries, materials, or trade flows in isolation, overlooking the dense upstream-to-downstream interdependencies that characterise modern supply chains. The authors call for stage-aware, multilateral coordination to ensure global supply chain resilience, and they specifically flag the limitations of purely national strategies.
This conclusion lands as a pointed challenge to the current architecture of Western critical minerals policy. Building on my analysis of China's layered mineral statecraft in June 2026, the structural mismatch between Western stockpile strategies and China's vertically integrated control is directly relevant here. The cobalt study makes the same structural point in quantitative terms: national stockpiling programs or production relocation efforts may lower risk for individual countries, but they can also shift vulnerabilities to other parts of the network rather than eliminating them. Improving resilience requires coordinated strategies that account for connections between upstream and downstream production stages. Focusing only on national interests without considering broader network relationships could unintentionally worsen instability across the global system.
The United States' current policy mix includes Project Vault, a US$12 billion critical mineral stockpile, more than US$30 billion in letters of interest and loans mobilised for supply chain projects over the past six months, a December 2025 strategic partnership with the DRC, and a 40% stake MoU in Glencore's two DRC copper and cobalt mines through the Orion Critical Mineral Consortium. The Inflation Reduction Act's FEOC rules create escalating pressure on automakers, with the critical minerals sourcing threshold rising from 60% in 2025 to 80% by 2027, making non-Chinese cobalt refining a commercial necessity for IRA credit eligibility. However, as LSE research economist Viet Nguyen-Tien told Energy Intelligence, subsidies for mines will not be sufficient by themselves; displacing or diversifying away from China requires sustained targeted support for midstream processing and refining, investment in skills and R&D, allied coordination on demand and offtake, and major pushes on recycling institutions, with competitiveness measured in years to decades rather than quarters.
The European position is materially weaker. The EU's Critical Raw Materials Act sets 2030 targets of at least 40% of critical raw material processing occurring domestically and no more than 65% of any strategic material sourced from a single third country. For cobalt, where China processes approximately 78% and the DRC mines approximately 73%, both thresholds are currently breached by wide margins. The December 2025 RESourceEU action plan is intended to accelerate investment, but Europe's non-Chinese cobalt refining capacity remains minimal, and the realistic timeline for building meaningful capacity is measured in years. Market participants at the time of the US-DRC strategic partnership announcement characterised the European position in stark terms: the EU is fighting over scraps and has demonstrated complacency relative to the pace of US and Chinese strategic positioning.
The indirect exposure finding in the study has direct relevance for procurement strategy. The avalanche network being four times denser than the visible physical trade network means manufacturers with no direct cobalt trade relationship with the DRC or China can still face severe downstream exposure through indirect supply chain dependencies. For procurement professionals and battery manufacturers, this finding underscores what DRC-based logistics and market data already confirm: spot market procurement strategies are no longer viable for cobalt-dependent manufacturers. Long-term offtake agreements at elevated price points have become the primary risk management tool for battery cell manufacturers across China, South Korea, Japan, and Europe.
Battery Chemistry Shifts and the Limits of LFP as a Structural Solution
One partial structural response to cobalt concentration risk has already been underway for several years in the form of battery chemistry diversification. Lithium iron phosphate batteries contain no cobalt and accounted for over 55% of EV batteries deployed globally in 2025, up from nearly 50% in 2024. LFP packs were more than 40% cheaper on average than NMC alternatives on a per-kWh basis in 2025. High-nickel chemistries such as NMC 811 rely on relatively small quantities of cobalt per cell. Together, these trends have meaningfully reduced cobalt's share of overall EV battery cost compared to the 2018 to 2022 period.
However, chemistry diversification does not dissolve cobalt's systemic risk profile, for several reasons. In the United States, superalloys accounted for approximately 51% of cobalt consumption in 2025, followed by chemical applications at roughly 25%, meaning cobalt dependency in defence and aerospace applications is structurally independent of EV battery chemistry trends. Globally, more than two-thirds of mined cobalt (71% in 2023 according to the Cobalt Institute's annual report) is used in lithium-ion batteries, and the EV sector accounts for approximately 40% of the global cobalt market. With global demand reaching approximately 219,600 tonnes in 2026 against a projected deficit of 10,700 tonnes, the market is structurally tight regardless of LFP's share gains.
The study's framework is deliberately chemistry-agnostic in its core insight. The avalanche network and robust-yet-fragile architecture it identifies are properties of the cobalt supply system's topology, not its end-use applications. A shift toward lower-cobalt chemistries reduces the magnitude of demand at risk but does not alter the cascade dynamics at refining and manufacturing bridges, which remain concentrated in China regardless of whether the downstream product is NMC 811, NMC 532, or cobalt sulphate for consumer electronics. The authors note explicitly that their framework could be applied to other materials essential for batteries and clean energy technologies, pointing toward a broader research programme for manganese, nickel, and lithium supply chains that may exhibit analogous topological vulnerabilities.
Conclusion: Quantified Fragility, Unresolved Architecture
The Ouyang et al. study published in Environmental Science and Ecotechnology (DOI: 10.1016/j.ese.2025.100654) provides the most rigorous quantitative foundation yet for a proposition that market participants and policymakers have been intuiting for several years: the global cobalt supply chain is not merely concentrated, it is structurally architected in a way that converts targeted shocks into nonlinear, cascading failures at a scale four times greater than visible physical trade flows suggest.
The three-layer concentration dynamic (DRC mining, Chinese refining, Chinese corporate control in DRC operations) means that the nodes the study identifies as highest-fragility are already under stress from deliberate policy action. The DRC quota system is, in the study's framing, a controlled experiment in what happens when a high-fragility upstream node restricts output: prices rise 160% from lows, hydroxide payables surge 350%, forecast deficits of 5,000 to 6,000 tonnes materialise in 2026 and 2027, and logistical failures at the DRC border compound upstream restriction into refinery feedstock tightness with a three-to-four-month transmission lag. The avalanche network is not hypothetical. Its density is being mapped in real time by the price discovery process.
The study's policy prescriptions, specifically stage-aware multilateral coordination and shared approaches to stockpiling, refining diversification, and trade restriction impact assessment, describe a governance architecture that does not currently exist at any meaningful scale. What exists instead is a combination of national stockpile programs, bilateral partnerships of variable depth, chemistry diversification by individual manufacturers, and IRA-style demand-side incentive structures that create pressure for supply chain change without the midstream and downstream processing infrastructure required to realise it. As the LSE's Viet Nguyen-Tien observed, that competitiveness will take years, likely decades to build.
The study's concluding observation is calibrated and precise: a successful low-carbon transition will depend not only on securing access to essential resources, but also on understanding and managing the complex global networks through which those resources flow. The cobalt market in mid-2026 is a case study in what happens when that network understanding lags behind both the physical concentration of supply and the policy sophistication of the actors controlling it. The avalanche network is four times denser than the trade data suggests. The institutional response remains a fraction of the density required to manage it.