Introduction

Mineral resources have long been a source of leverage and conflict between competing nations (Barbera 1982; Dumett 1985). The current focus on nonfuel minerals, especially critical minerals that enable advanced technologies, was precipitated by the Chinese government’s abrupt reductions of rare earth element (REE) export quotas from 2009 through 2012 (Eggert 2011; Gambogi 2014). Chinese exports were virtually the only supply source at the time because roughly 90% of world REE mine production occurred in China (Eggert 2011; Gambogi 2014). In 2010 Chinese customs officials reportedly blocked REE exports to Japan during a diplomatic dispute (Bradsher 2010; Humphreys 2015). This sparked a global urgency to stockpile REE supplies, which are indispensable for the manufacture of technologies from rechargeable batteries for hybrid electric vehicles, to permanent magnets for wind turbines, to jet engines for commercial and military aircraft (Eggert 2011; Gambogi 2014). Besides rare earth elements, perhaps no case has received more attention for its supply risks than cobalt (Andersson and Råde 2001; Nansai et al. 2014; Helbig et al. 2018; Gulley et al. 2019; Sun et al. 2019)—which is rated in the top 3, 4, and 10 most critical minerals by the United States, China, and the European Union, respectively (Blengini et al. 2020; Nassar and Fortier 2021; Yan and Wang 2021).

Due to its industrial and military applications, cobalt became a critical mineral to the defense industrial base in World War I and the United States Congress mandated that it be stockpiled in 1946 (Barbera 1982). Prior to the year 2000, cobalt was mostly known for its strategic applications including REE permanent magnets, jet engines, missile propulsion systems, and cutting/wear tools that enable modern construction, manufacturing, oil drilling, and mining (Shedd 2002; Gulley 2022). Since 2000, cobalt has become better known by its use in lithium-ion battery manufacturing. This consumption grew more than 43-fold between 2000 and 2022, with 89% of this growth occurring in China (Shedd 2002; Darton 2022; Gulley 2023).

More recently cobalt has become associated with child labor at artisanal cobalt mine sites (Amnesty International 2016; Associated Press 2019; Niarchos 2021), with production from this source estimated to represent 2% of world cobalt mine production in 2020 (Gulley 2023). Research on cobalt since 2000 has primarily focused on potential mine production disruptions (Olivetti et al. 2017; Alves Dias et al. 2018; Nassar et al. 2020), the possibility of production shortfalls (Andersson and Råde 2001; Olivetti et al. 2017; Alves Dias et al. 2018; Xu et al. 2020), potential reserve depletion (Andersson and Råde 2001; Alves Dias et al. 2018; Xu et al. 2020), and humanitarian issues related to artisanal mining (OECD 2013; Amnesty International 2016; EGC 2021). However, relatively little attention has been dedicated to the main factor that disrupted REE markets in 2010—namely Chinese control of production and the willingness to leverage this control during diplomatic and trade disputes (Ericsson et al. 2020, 2023).

This analysis therefore estimates the control that Chinese firms hold over the world’s production, consumption, and stocks of each material in the cobalt supply chain from 2000 through 2022. Control over the minerals sector in China is depicted as being highly centralized (Ericsson and Löf 2011; Riddle et al. 2015; Brown and Eggert 2018; Gulley et al. 2019). Chinese control is therefore estimated by summing production in China, production from Chinese majority-controlled foreign assets, and known offtake agreements between Chinese and non-Chinese firms. This analysis provides 23 years of granular data regarding cobalt mine production, refinery production, recycling, battery-grade chemical production, battery cathode precursor production, battery cathode production, consumption, production loss percentages, production loss quantities, annual surpluses, and Chinese trade.

Individual Chinese foreign cobalt investments are detailed in Supplementary Table 1. Numerical production data from 2000 through 2022 are presented by material and country in Supplementary Tables 2 (mining), 3 (refining), 4 (secondary production from recycling), and 5 (battery-grade refining, as well as battery cathode precursor, and battery cathode production). Production details and data sources are found in the Data and Methods section, as well as Supplementary Discussion 1. Numerical data regarding consumption for manufacturing by application are shown in Supplementary Table 6, while consumption details and data sources appear in the Data and Methods section, as well as Supplementary Discussion 2.

In terms of world production, the results show that China’s control over the cobalt supply chain increased from 2000 through 2022 from 0 to 62% for chemical refinery feedstock and from 10 to 95% for refined cobalt chemicals (Supplementary Tables 2 and 3, respectively). In the lithium-cobalt-oxide (LCO) cathode supply chain, China’s control increased from 2000 through 2022 from 0 to 92% for battery-grade cobalt oxides and from 0 to 84% for LCO cathode materials (Supplementary Table 5). Similarly, in the nickel–cobalt-manganese (NCM) cathode supply chain, China’s control increased from 2008 through 2022 from 0 to 85% for battery-grade cobalt sulphate, from 0 to 91% for NCM precursor materials (Supplementary Table 5). China’s monopoly over these materials coincides with an increasing focus by China’s government on supply chokepoints, “to accelerate the leapfrog development of the new energy industry” (PRC National Development and Reform Commission 2020). China has recently demonstrated its focus on emerging technology by its export controls on gallium, germanium, graphite, and rare earth elements which enable technologies for integrated circuits, solar cells, electronic warfare, lithium-ion batteries, and permanent magnets (Tan 2023).

The focus of Chinese firms on critical minerals for emerging strategic industries is highlighted by a previous study, which estimated Chinese control to be just 3.4% of the value of all global mine production, 6.4% for all of Africa, and 24% for all of the DRC in 2018 (Ericsson et al. 2020). In contrast, for the year 2018, Chinese control over the value of African copper and cobalt production was 28% and 41%, respectively (Ericsson et al. 2020). This result is rather close to China’s 37% control of production estimated in this study and 56% when offtake agreements are included, which Ericsson et al. 2020 did not.

Cobalt supply chain stages and cobalt materials

The first step in estimating Chinese control is to delineate the predominant production pathways that move cobalt materials out of cobalt-containing reserves to the manufacturing applications that consume refined cobalt. Based on academic literature in the field of extractive metallurgy (Crundwell et al. 2011, 2020; Schmidt et al. 2016), government reports (Shedd 2002), and industry reports (Darton 2022), this analysis provides a detailed examination of the cobalt supply chain—especially the portion that feeds battery cathode production. Each reserve type, such as copper-cobalt versus nickel-copper-cobalt, is often suitable for the production of several specific cobalt materials, such as crude cobalt hydroxide versus nickel-copper-cobalt matte (Crundwell et al. 2011, 2020; Schmidt et al. 2016). Given that these cobalt mine materials serve as feedstock for cobalt refineries, they are also referred to here as refinery feedstock. The disaggregation of mine material production is primarily achieved by separating facility-level production into categories that match the type of reserve (Cu-Co, Ni-Co laterite, Ni-Co sulfide, PGE-Co, or Co-As) with the type of mine materials produced (Shedd 2002; Darton 2022).

Each cobalt mine material is then mostly consumed to produce a specific type of refined cobalt material (e.g., commercial-grade cobalt chemicals versus cobalt metal) (Crundwell et al. 2011, 2020; Schmidt et al. 2016). The disaggregation of refinery production is achieved by matching refinery feedstock input materials with the type of refinery production (chemical or metal). This matching is primarily achieved at the facility-level, but country-level sources are also used when facility-level information are not available. Refined cobalt products can be divided into chemicals, metals, coarse powder, and fine powder (Shedd 2002; Darton 2022). Each refined cobalt material is then chiefly consumed by a specific application (e.g., battery-grade cobalt chemicals versus steel alloys), which is the first-use for manufacturing of cobalt containing products (Crundwell et al. 2011, 2020; Schmidt et al. 2016). Due to the surplus of commercial-grade cobalt chemicals and the relatively higher price of cobalt metal, 2023 was an anomaly because metal producers used unusually large amounts of crude cobalt hydroxide to produce metal (Darton 2024).

In this way, one may characterize what manufacturing application is likely to consume the production of each reserve type. It also illuminates how the production path of each reserve type, mined material, and refined material changes to satisfy changes in cobalt consumption for manufacturing. For example, prior to 2000 copper-cobalt deposits predominantly produced materials that largely fed cobalt metal refining, which in turn mostly fed into superalloy manufacturing (Shedd 2002; Gulley 2022). As battery demand for cobalt chemicals increased in the 2000’s cobalt refining switched from mostly metal to mostly chemicals (Darton 2009). In the late 2000’s, the government of the Democratic Republic of the Congo (DRC) required cobalt mined materials to be processed before they could be exported (Darton 2009). The production of crude cobalt hydroxide emerged as the lowest cost inputs for commercial-grade cobalt chemical refining (Darton 2009). By 2022 output from copper-cobalt deposits flowed mainly to commercial-grade chemical refining which principally fed the production of battery-grade cobalt chemicals (Darton 2021; Gulley 2022). Figure 1 presents the simplified structure of the world cobalt supply chain with mined, intermediate, and refined materials, as well as consumption for manufacturing, for the year 2022.

Fig. 1
figure 1

World cobalt supply chain for the year 2022. White boxes are production processes. Solid-color boxes are material outputs/inputs whose width corresponds to world production. The dark colored portion of the dotted boxes represent production under Chinese control. The light colored portion of the dotted boxes represent all other production. As, Co, Cu, LCO, NCM, NCA, Ni, and PGE refer to arsenic, cobalt, copper, lithium-cobalt oxide, nickel–cobalt-manganese, nickel–cobalt-aluminum, nickel, and platinum-group elements, respectively. Numerical data and sources are provided in Supplementary Tables 2, 3, 5, and 6

Full size image

The left side of Fig. 1 shows the supply chain predominantly for cobalt chemical refining for battery manufacturing. Here, the mining stage is dominated by the production of crude cobalt hydroxide from copper-cobalt (Cu-Co) reserves. In 2022 crude cobalt hydroxide was produced almost exclusively in the DRC (Darton 2020). Crude hydroxide and, to a lesser extent nickel–cobalt mixed hydroxide precipitate (MHP) from nickel-copper-cobalt (Ni-Co) reserves, were the principal feed for cobalt chemical refining (Crundwell et al. 2011, 2020; Schmidt et al. 2016). Refined cobalt chemicals then served as the primary feed for the cobalt-containing battery cathode supply chain (Crundwell et al. 2011, 2020; Schmidt et al. 2016). In the battery cathode supply chain, Fig. 1 shows refined cobalt chemicals fed the production of battery-grade cobalt oxide and battery-grade cobalt sulphate (Darton 2022). Battery-grade oxide then fed lithium-cobalt-oxide (LCO) precursor production, which fed LCO cathode production (Darton 2022). Battery-grade sulphate fed both nickel–cobalt-manganese (NCM) and nickel–cobalt-aluminum (NCA) precursor production, which then fed NCM and NCA cathode production, respectively (Darton 2022).

The right side of Fig. 1 shows the supply chain predominantly for cobalt metal refining. Here, the production of metal refinery feedstock (i.e., metal feed) was consumed for metal refining, which then predominantly fed cobalt metal consuming industries such as superalloys and permanent magnets (Crundwell et al. 2011, 2020; Schmidt et al. 2016). A more granular presentation of the cobalt supply chain is provided in the Data and Methods section, as well as Supplementary Fig. 1 in Supplementary Discussion 2. Additionally, Supplementary Discussion 3 and Supplementary Fig. 2 depict a supply chain comparison between 2000, 2010, and 2019.

Data and methods

Cobalt production data

Chinese foreign cobalt investments, as shown in Supplementary Table 1, have been almost exclusively targeted at mine and intermediate (i.e., refinery feedstock) producing assets in the DRC. Mine and intermediate production data are primarily collected at the mine- (or facility-) level from governmental (Shedd 2002), industry (Darton 2022) and Non-Governmental Organization (NGO) (EITI 2021) reports. When mine-level data are not available, production is collected at the country-level from government reports (Shedd 2002). Production data for mined and intermediate products by ore type, refinery feedstock type, ownership type, and country are presented in Supplementary Table 2.

Chemical and metal refinery data are also collected primarily at the facility level (Shedd 2002; Darton 2022). These data provide country-level production estimates for chemical refining and metal refining separately. Refinery production data are presented in Supplementary Table 3. In contrast to the production data available for the materials above, the production of battery cathode precursors and battery-grade cobalt chemicals is estimated from cathode production data and average reported loss rates (see Supplementary Table 5 for data sources). For China, net trade of battery cathode, precursor, and battery-grade chemicals is also employed to estimate domestic production. Production estimates and data sources for battery-grade chemicals, cathode precursor materials, and battery cathode materials are presented in Supplementary Table 5. Estimates for secondary production (i.e., recycling) by application are provided in Supplementary Table 4. See Supplementary Discussion 1 for data sources and estimation methods.

Chinese control by material

A main goal of this work is to quantify the extent to which Chinese firms control global supply chains for specific cobalt-consuming applications such as refined chemicals for lithium-ion batteries versus refined metals for rare earth element permanent magnets or superalloys for jet engine turbines. Now that the supply chain has been disaggregated and production data have been collected and estimated for each material, Chinese control is estimated. Here, Chinese control is calculated as the sum of China’s domestic production plus the production from foreign operations where Chinese firms hold a controlling equity share (> 50% ownership) plus known offtake agreements purchased by Chinese firms. All domestic mining is assumed to be under Chinese control due to the highly centralized nature of China’s mining industry (Gulley 2023). A controlling equity share provides Chinese firms with control over key operational decisions. As such, this analysis assigns all production to Chinese control, which provides a more accurate estimate of Chinese control than previous analyses that equated production control with percent of equity share (Gulley et al. 2019). Offtake agreements are included because they serve as another method of securing mineral supplies. To inform how Chinese control has evolved over time for each cobalt material, it is estimated from 2000 through 2022. For a discussion of how China’s investments relate to: centralized planning of China’s mineral sector, China’s 2000 “Going Out Strategy”, China’s 2007 “minerals-for-infrastructure agreement”, the Chinese government’s support of outward foreign direct investment (OFDI) in the mining and related sectors, the role of state-owned versus private firms, and development plans for strategic emerging industries such as electric vehicles (i.e. Five Year Plans, Made in China 2025, and the Belt and Road Initiative)—see Gulley et al. 2019.

Cobalt stock changes and stocks of mined and refined materials

Since 2000 there have been concerns that cobalt consumption would soon exceed production. To illustrate this issue, this analysis evaluates the annual balance between primary plus secondary cobalt production and consumption to calculate annual surpluses/deficits for the world and in China for chemical refinery feedstock, metal refinery feedstock, refined chemicals, and refined metals from 2000 through 2022. The major data required are primary (i.e., mined, intermediate or refinery) production, secondary production (i.e., recycling), losses during production, and material consumption (i.e., refinery production for refinery feedstocks or manufacturing industry consumption of refined products). See Supplementary Table 6 for data on cobalt consumption by manufacturing applications. Notably, consumption for manufacturing (or for the subsequent production stage) is different than demand. Demand equals the production of the feedstock materials plus stock changes of those materials. Similarly, production is different than supply which includes production, stock changes and trade (when country-specific). By comparing production and consumption, stock changes can be estimated—whereas comparing supply and demand obscures stock changes.

World annual surpluses are calculated by subtracting each year’s consumption and losses from production, while stocks are calculated by summing annual surpluses from prior years. See Supplementary Discussion 4 and Supplementary Tables 7 and 8 for the data and methods behind cobalt production losses. See the Supplementary Discussion 5 for a more detailed discussion of the methodology, the data sources relied upon, and the surplus and stock estimates from 2000 through 2022. See Supplementary Table 9 for detailed numerical estimates of world surpluses and stocks. Chinese annual surpluses and stocks are calculated using the same method as for the world, except that net imports of each material are added to the Chinese production. See Supplementary Table 10 for detailed numerical estimates of China’s surpluses and stocks.

Estimating China’s cobalt material trade

As discussed, China’s net trade is often calculated to estimate the supply (here defined as domestic production plus net imports) available for domestic consumption for manufacturing or the subsequent supply chain stage. Trade estimates are calculated by multiplying the reported cobalt content percentages by trade data for the relevant harmonized tariff system (HTS) and tariff line trade codes. The raw trade data are obtained from the data vendor IHS Global Trade Atlas (GTA) based on the primary quantity, which is the gross weight (IHS Markit 2022). Reported cobalt content percentages are obtained from academic (Crundwell et al. 2011, 2020; Harper et al. 2012a; Schmidt et al. 2016; Darton 2022), industry (Xu 2017; Darton 2020, 2021), and NGO (Extractive Industries Transparency Initiative 2017) publications. Reported contents for cobalt chemicals are close to (but lower than) stoichiometric cobalt contents. Imports are subtracted from exports to estimate net imports. Net imports represent the cobalt content of imports minus the cobalt content of exports. Positive numbers represent net imports and negative numbers represent net exports. To illuminate the magnitude of China’s imports of mine and intermediate cobalt materials, in relation to total world trade, the gross weight of China’s imports is divided by the gross weight of all countries’ imports for mine materials as well as intermediate materials. When analyzing international trade data, reported import quantities are typically more reliable than reported export quantities. This is largely because the incentive to collect import duties encourages importing nations to track incoming goods and quantities more meticulously. China’s exports, therefore, are estimated by summing the quantities of each countries’ reported imports from China. Trade for the relevant HTS code is multiplied by the relevant reported quantity. See Supplementary Discussion 6 for details on calculations, data sources and methods. See Supplementary Table 11 for numerical estimates.

Results and discussion

China’s increasing control over the cobalt supply chain

No government, perhaps, has created more policies to advance domestic manufacturing capabilities and secure foreign mineral resources than China (Gulley et al. 2018, 2019; PRC National Development and Reform Commission 2020). To alleviate mineral supply chokepoints for China’s strategic emerging industries such as electric vehicles, China’s government has supported OFDI in mineral and infrastructure projects through a series of outward facing policies including the Going Out Strategy, the Belt and Road Initiative, and the 2007 minerals-for-infrastructure agreement with the DRC (Gulley et al. 2019; Gulley 2022).

As mentioned above, Chinese control is estimated by summing production in China, production from Chinese controlled foreign assets, and known offtake agreements between Chinese and non-Chinese firms. See the Data and Methods section, as well as Supplementary Discussion 3 for details. In 2022, Chinese control over mine and intermediate production (i.e., refinery feedstock) came almost entirely from OFDI in cobalt producing facilities in the DRC, with the remainder from developments of MHP mines and processing facilities in Indonesia (Darton 2022). Chinese control in refinery, battery-grade chemical, precursor and cathode production is limited almost exclusively to facilities within China (Gulley et al. 2019). From 2000 through 2022, Figures 2a and b, respectively, display that China’s control increased from 0 to 62% for cobalt chemical refinery feedstock and from 10 to 95% for cobalt chemical refinery production. In comparison, from 2000 through 2022 China’s control only increased from 0 to 5% for cobalt metal refinery feedstock and from 0 to 26% for cobalt metal refinery production (Fig. 2a and b, respectively). It is worth noting that preliminary estimates of China’s control over world cobalt metal production increased to 43% in 2023 (Darton 2024). This increase was largely in the form of aerospace-grade metal, suitable for jet engine superalloys for commercial and military aviation (Darton 2024). China’s material stockpiling authority, the State Reserve Bureau (SRB), purchased 93% of this additional aerospace-grade metal (Darton 2024).

Fig. 2
figure 2

World cobalt material production and first-use consumption for manufacturing by Chinese control (domestic plus foreign owned plus offtake agreements) and other or rest of world (left axis, thousand tons contained cobalt per year), as well as the percentage of world production under Chinese control (right axis, percentage). a, Mine and intermediate production by refinery feedstock type (chemical or metal). b, Refinery production by type. c, Battery-grade (BG) chemical production by type. d, Battery cathode precursor production by type (BG oxide is replicated here because LCO does not have an equivalent precursor material). e, Battery cathode production by type. f, Cobalt-consuming manufacturing applications by type. Data for Chinese NCM battery-grade chemical, cathode precursor, and cathode production are from 2009 through 2022 to reflect the beginning of NCM production in China. For the same reason Chinese NCA data do not begin until 2016. BG, LCO, NCM, and NCA refer to battery-grade, lithium-cobalt oxide, nickel–cobalt-manganese, and nickel–cobalt-aluminum, respectively. Numerical data and sources are provided in Supplementary Tables 2 through 6

Full size image

In the year 2000, chemical refinery feedstock production was dominated by non-Chinese owned cobalt concentrate production in the DRC (Supplementary Table 2). This began to change following the 2007 minerals-for-infrastructure agreement as Chinese owned crude cobalt hydroxide production started to increase. After purchasing majority shares of Tenke Fungurume in 2016 and 2017 (Darton 2018), Chinese controlled production was over half of DRC and world chemical refinery production (Supplementary Tables 1 and 2). In the year 2000, chemical refinery production was dominated by Finland, Canada, Australia, and Belgium (Supplementary Table 3). Beginning in 2000, China rapidly expanding chemical refinery production so that by 2008, refined more cobalt chemicals than the rest of the world (Supplementary Table 3). This expansion continues with China refining 95% of the world’s cobalt chemicals in 2022 (Supplementary Table 3). Although China’s metal production is less dominant, the country overtook the largest producers of DRC, Zambia, and Russia by the year 2005 (Supplementary Table 3). This dominance also persists with China producing twice as much cobalt metal as its nearest competitor, Canada, in 2022 (Supplementary Table 3).

Figure 2c, d and e, respectively, show that within the LCO supply chain, from 2000 through 2022 China’s control increased from 0 to 92% for battery-grade tricobalt tetroxide production and from 0 to 84% for LCO cathode production. As for the NCM and NCA supply chains, from 2008 (the first year that data are available) through 2022 China’s control increased from 0 to 85% for battery-grade cobalt sulphate production, from 0 to 91% for NCM precursor production and from 0 to 73% for NCM cathode production (Fig. 2c, d, and e, respectively). During the same period, China’s control over NCA precursor and NCA cathode production increased from 0 to 19% and from 0 to 8%, respectively (Fig. 2d and e). These results indicate that cobalt investments from Chinese firms have targeted the portion of the cobalt supply chain that feeds battery cathode production.

This conclusion is supported by the 109-fold increase in China’s cobalt chemical refinery production from 2000 through 2022. Further, as world cobalt consumption for battery manufacturing grew 43-fold during the same period, 89% of that growth occurred in China. Figure 2f reveals that, from 2000 through 2022, China’s share of world cobalt consumption for manufacturing grew in every industrial application: batteries, 3% to 87%; magnets, 0% to 63%; hardmetals and cemented carbides, 23% to 58%; ceramics and pigments, 15% to 35%, catalysts, 7% to 34%; and superalloys, 0% to 21%. See Supplementary Table 6 for numerical estimates and Supplementary Discussion 2D for data sources. Regarding secondary production (i.e., recycling), from 2000 through 2022 China’s control is also estimated to have increased in all industrial applications: batteries, 0% to 75%; catalysts, 0% to 67%; hardmetals and cemented carbides, 11% to 67%; and superalloys plus magnets, 13% to 50%. These numbers are not certain, however, as the quantity of recycled material inputs for world cobalt refining or for manufacturing of cobalt-containing products is poorly understood and not reported reliably or consistently. Numerical secondary production data are presented in Supplementary Table 4, while details and data sources can be found in the Data and Methods section, as well as Supplementary Discussion 1D.

Chinese control over cobalt stocks

Predictions that cobalt availability will likely constrain world deployment of advanced batteries have been discussed and forecast since the year 2000 (Andersson and Råde 2001; Olivetti et al. 2017; Alves Dias et al. 2018; Valero et al. 2018; Lèbre et al. 2020; Xu et al. 2020). These assessments are addressed here by evaluating the annual balance (i.e., surplus or deficit stock change) between cobalt production and consumption at each supply chain stage—either for the next production process or for first-uses for manufacturing. World annual stock changes are calculated by starting with each year’s production and then subtracting consumption and production losses. Cumulative stocks are calculated by summing annual stock changes from previous years. Chinese annual stock changes and cumulative stocks are calculated using the same method as for the world, except that net imports of each material are added to Chinese production. Annual stock changes are estimated for the world and in China for chemical refinery feedstock, metal refinery feedstock, refined chemicals, and refined metals from 2000 through 2022.

Each production process discussed above and presented in Figs. 1 and 2 incurs some percentage loss of cobalt content. Production yields and losses by major ore type, production process, and final refined product are aggregated from the literature, presented in Supplementary Table 7, and discussed in Supplementary Discussion 4. Supplementary Table 8 multiplies these loss estimates by the production in Supplementary Tables 2 and 3. This provides estimates of the quantity of contained cobalt lost during each production route. Stock change and stock calculation methods are described in the Data and Methods section, as well as Supplementary Discussion 5. Supplementary Tables 9 and 10 present stock changes and stocks estimates for the world and China, respectively. Supplementary Discussion 6 provides a description of trade calculations for China. Numerical estimates of China’s imports, exports, and net imports are presented by material form in Supplementary Table 11.

The results show that world production of cobalt refinery feedstock and cobalt refined materials has chiefly exceeded consumption for the last 20 years. Figure 3 summarizes this result by presenting the cobalt content of refinery feedstock (blue dotted line), refined materials (purple dashed line), and consumption for manufacturing (orange area for batteries and green area for all other). Each year, the distance from the blue dotted line to the purple dashed line represents the stock change for refinery feedstock (i.e., mined, intermediate, and recycled materials), while the area between the two lines represents the cumulative stock. For stock changes and cumulative stocks of refined materials, the same applies to the difference between the purple dashed line and the green plus orange areas. To show market responses, the refined cobalt price is presented as a grey line on the right axis (Kelly and Matos 2020; Gulley 2022; U.S. Bureau of Labor Statistics 2022; Shedd 2023). Cobalt market events, production-consumption changes, and trends that may have impacted the world refined cobalt market are also provided.

Fig. 3
figure 3

World cobalt consumption (battery in orange and all other in light green stacked area graph), world cobalt refinery production (dashed purple line), and loss-adjusted refinery feedstock production (dotted blue line) are presented in metric tons of cobalt content (left axis) for the years 2000 through 2022. The annual average refined cobalt metal price, deflated to 2020 US dollars (2020$) by the producer price index, is the grey line (right axis). EV refers to electric vehicles. Numerical consumption, refinery production, and feedstock production data and sources are provided in Supplementary Tables 2, 3, and 6 respectively. Price data are from (Kelly and Matos 2020; Gulley 2022; U.S. Bureau of Labor Statistics 2022; Shedd 2023)

Full size image

Estimated annual stock changes and stocks are presented in Fig. 4 for the world and China, by material, for each year from 2000 through 2022. In Fig. 4, the term ‘supply’ for China equals production plus net imports only. Since 2000, there were no annual deficits estimated for chemical refinery feedstock. These annual surpluses are estimated to have resulted in 2022 cumulative world stocks equal to 10 months of chemical refinery consumption (Fig. 4a). Since 2000 there were seven world annual deficits estimated for refined cobalt chemicals (2006, 2007, 2012, 2016, and 2020–2022), resulting in estimated 2022 cumulative world stocks equal to 7 months of refined chemical consumption (Fig. 4c). Although these stocks are substantial, annual deficits from 2020 through 2022 indicate that cobalt chemical refining is not keeping pace with reported cobalt-containing battery cathode production.

Fig. 4
figure 4

Annual surpluses and stocks for the world and China from 2000 through 2022 (thousand tons contained cobalt per year). a, World chemical refinery feedstock. b, World metal refinery feedstock. c, World refined chemicals. d, World refined metals. e, China chemical refinery feedstock. f, China metal refinery feedstock. g, China refined chemicals. h, China refined metals. Numerical data and sources for a, b, c, and d are provided in Supplementary Table 9. Those for e, f, g, and h are provided in Supplementary Table 10

Full size image

World metal refinery feedstock annual deficits were more common than those for chemical refinery feedstocks (2005, 2016–2019, and 2022). Nonetheless, estimated 2022 cumulative world stocks were equal to 9 months of metal refining consumption. Refined metal deficits were estimated for the years 2000, 2010, 2015–2022, with estimated 2022 cumulative world stocks equal to 4 months of metal consumption (Fig. 4d). The prevalence of annual deficits in the cobalt metal supply chain since 2016 may indicate that, as the cobalt supply chain shifts towards chemical consumption, there is less material available for metal refining and metal consuming industries.

Of world stocks in 2022 China held 31% of world chemical refinery feedstocks (Fig. 4e versus Fig. 4a), 35% of refined chemical stocks (Fig. 4g versus Fig. 4c), 6% of metal refinery feedstock (Fig. 4f versus Fig. 4b), and 62% of refined metal stocks (Fig. 4g versus Fig. 4c)—resulting in 2022 Chinese stocks equal to four, three, two, and seven months of consumption, respectively. China’s relatively large chemical refinery feedstock and refined chemical stocks reflect China’s large consumption of cobalt materials for batteries. Similarly, China’s relatively small metal refinery feedstocks reflect China’s smaller production of cobalt metals. However, the large stocks of refined metal may be explained by the fact that countries have stockpiled cobalt metal since World War II in strategic reserves (Gulley 2022). From 2014 through 2022, the Chinese State Reserves Bureau (SRB) has reportedly purchased 10,300 t of cobalt metal for its own strategic reserves (Gulley et al. 2019; Darton 2021).  Preliminary estimates indicate that the SRB stockpiled an additional 8,700 t of aero-space grade metal in 2023 (Darton 2024).

Key uncertainties

The estimates of Chinese control for cobalt battery-grade chemicals and cathode precursors represent the main uncertainty in the analysis. This uncertainty stems from the reliance on estimates of global and Chinese cathode production by type reported by industry sources, as well as trade flows of battery-grade chemicals and precursors reported by China Customs. An example of this uncertainty is LCO which does not have its own trade code. As a result this analysis relies on LCO imports and exports reported by Chinese industry (Xu 2017) and academia (Chen et al. 2019). Estimates for China’s LCO trade were not available from 2017 through 2022. Further, reported battery cathode production data by cathode type for China and the world vary considerably between available information sources (Roskill 2007, 2019, 2021a; Benchmark Mineral Intelligence 2021; Darton 2021; Xu et al. 2021). This introduces uncertainty into this analysis and provides a key avenue of future research for battery material studies.

Additional uncertainty relates to data on secondary production from recycling. The quantity of recycled material inputs for world cobalt refining or for manufacturing of cobalt-containing products is poorly understood and not reported reliably or consistently (Shedd 1993; Harper et al. 2012b; Roskill 2019; Darton 2021). In the absence of clearly reported quantities of recycled inputs or secondary production, these quantities are either estimated based on assumed product lifetimes and recycling rates (Harper et al. 2012b; Zeng and Li 2015; Sun et al. 2019) or on information gleaned from market participants (Darton 2019, 2020, 2021; Roskill 2021b). Supplementary Table 4 presents world and Chinese secondary production from batteries (Darton 2018, 2019, 2020, 2021; Sun et al. 2019), catalysts (Darton 2018, 2019, 2020, 2021; Roskill 2019; Sun et al. 2019), cemented carbides (Darton 2018, 2019, 2020, 2021; Sun et al. 2019; Roskill 2021b), and superalloys and magnets (Darton 2018, 2019, 2020, 2021; Sun et al. 2019; Roskill 2021b).

World secondary production data were estimated by category by the sources cited. Chinese secondary production was not reported by end-of-life products so the percentage of world secondary production from each end-of-life product for each year was assumed to be the same for China. One key caveat is that the estimated cobalt production from spent batteries in 2022 was over 4.5 times larger than that of 2015. This may be due to overreporting of Chinese battery recycling production, given that there are only a handful of battery recycling facilities.

Examples of Chinese control over other critical minerals for China’s strategic emerging industries

Cobalt is not alone with respect to China’s increasing dominance over critical mineral supply chains for strategic emerging industries. China’s dominance of several critical minerals is readily apparent from China’s share of world production. China’s share of gallium production, for example, increased from 0% in 2000, to 34% in 2005, to 36% in 2010, to 94% in 2015, to 97% in 2022. Germanium followed a similar path to gallium. While starting from a higher level, Chinese control over bismuth (42% in 2000) and graphite (51% in 2000) was 82% and 72%, respectively, in 2022. However, these numbers bely China’s control over specific commodity forms such as an estimated 99% of spherical graphite for batteries (Benchmark 2023). See Supplementary Table 12 for details.

Conclusion

Since 2000, world surpluses have been the norm for cobalt mined and refined materials—even as cobalt consumption for battery manufacturing has grown 43-fold. Deficits in refined cobalt materials for the years 2020 through 2022, especially for refined chemicals in China, are likely the result of drawdowns of large 2019 stock levels. Accumulation of China’s large stocks of refined chemicals may have been due to the relative lack of pressure on Chinese firms to adjust production when prices decline (Darton 2022). Although world stocks and production may continue to exceed world consumption, international competition and geopolitical tensions may play an important role in determining which countries have sufficient access to cobalt supplies to sustain their manufacturing industries. This analysis reveals that, without exception, Chinese control has increased at every cobalt supply chain stage for production, consumption, and recycling from 2000 through 2022. This is especially true for the refined chemicals, battery-grade chemicals, and battery-cathode precursor materials where Chinese control has increased to monopoly levels.

As was the case during the 2010 disruption of rare earth element markets, Chinese exports are virtually the only supply source of the refined cobalt chemicals and cobalt-containing battery-cathode precursors. These supplies could become unavailable either due to Chinese firms consuming all the Chinese production or due to the withholding of exports during an international dispute. The speed and level to which Chinese firms can gain control over mineral supplies for China’s strategic emerging industries has seen similar efforts for other critical mineral supplies including rare earth elements, gallium, germanium, graphite, lithium, bismuth. As a result, systematic evaluation of Chinese control may need to be included in current methods of evaluating competition for mineral supplies (Gulley et al. 2018).

It is worthwhile to mention a few of the many mitigation strategies that could be employed by the private sector, governments, and public/private partnerships. First, firms could build chemical refining facilities. To supply these facilities, firms could purchase chemical refining feedstock from the open market, arrange offtake agreements, purchase controlling shares of chemical refining feedstock mines/companies, or develop new mining or processing capacity (Darton 2021). Second, firms could pursue the generally more expensive route of securing cobalt metal supplies to dissolve for production of battery grade chemicals. Chinese firms only control 5% of metal refinery feedstock and 26% of metal refinery production. Volkswagen, for example, chose this route and secured an offtake agreement with a Moroccan metal producer (Darton 2021). Third, firms could form joint ventures with Chinese firms to secure an offtake agreement or construct new chemical refining capacity outside China. South Korean firm LG Chem, for example, has created a joint venture with Chinese firm Huayou Cobalt to build one battery material plant in South Korea and another in China (Darton 2022). Fourth, firms could focus on battery cathode chemistries that do not contain cobalt—as long as the non-cobalt-containing batteries can meet the performance requirements demanded by consumers.

Stakeholders in critical material supply chains should conduct more detailed analyses of mineral supply chains that evaluate each material at each supply chain stage. For example, indirect import reliance—in the form of imports from adversarial countries that pass through friendly intermediary countries—may reveal undiscovered supply risks. Research evaluating Chinese control over other critical minerals that feed China’s strategic emerging industries (i.e., lithium, nickel, graphite, or manganese in lithium-ion batteries) has become a priority for some governments (Trump 2020) and may warrant future investigation.