The Small Markets That Became Strategic

Five product families that barely registered in world trade a generation ago now sit underneath the energy transition and the AI hardware build-out: lithium compounds, the cobalt chain, graphite, copper and nickel ores, and rare earth metals. Bundled together as an editorial “critical battery minerals” group, their combined world export value rose from $9.8B in 2000 to $124.1B in 2024 in this BACI extract. That total is dominated by one line: copper and nickel ores alone were $108.9B in 2024, or 87.7 percent of the group, so the headline figure is really a copper-and-nickel story with a battery-metals tail.

The smaller markets are where the strategic tension lives. Lithium compounds went from $0.1B to $6.4B over the same window — a roughly 49.6-fold increase — while cobalt reached $5.7B and graphite $2.7B. None of these is large next to oil or semiconductors, but each is a chokepoint input, and each remains concentrated in a handful of exporters. This article measures the size, the growth, and the concentration of those markets, and separates how much of the growth is volume from how much is price.

Two cautions sit next to every number here. The grouping is editorial, not a BACI category; and commodity price cycles can move nominal trade value as much as physical flows do — a point the final section makes concrete.

On a log scale, the trajectories separate cleanly. Lithium compounds are the steepest line: from $0.1B in 2000 to $6.4B in 2024, a 49.6-fold rise that tracks the arrival of lithium-ion cells in phones, then laptops, then vehicles and grid storage. Copper and nickel ores grew about 15.4-fold to $108.9B, the largest absolute mover, reflecting both electrification demand and a long price up-cycle. Graphite (to $2.7B) and the cobalt chain (to $5.7B) grew more modestly in value terms but remain central to anode and cathode chemistry.

Rare earth metals are the smallest line and the most counter-intuitive. The narrow HS92 code used here, 280530, captures only $0.5B of trade in 2024 — a reminder that most rare-earth value travels embedded in magnets, oxides, and finished components under other codes, not as the metals themselves. The log view is deliberate: on a linear axis copper and nickel ores would flatten every other series into the baseline, hiding the fact that lithium and graphite each grew faster in percentage terms. Reading growth on a log scale is the only way to see five markets of very different sizes in one frame.

These are nominal US-dollar values, not inflation-adjusted, and they mix price and quantity. The size ranking is robust; the growth rates should be read alongside the price caveat below.

Concentration is the defining feature. In 2024, Chile supplied 48.7 percent of world lithium-compound exports and China 31.9 percent, with Argentina a distant third at 10.6 percent — three countries account for the overwhelming majority. The cobalt chain is even more lopsided: the Democratic Republic of the Congo alone was 54.6 percent of world exports. Graphite shows China at 50.1 percent, just over half the market on its own.

For the headline critical-minerals group as a whole, the leaders are Chile at 26.9 percent (about $33.4B), Peru at 16.6 percent, and Indonesia at 6.6 percent — a ranking driven by their copper, and in Indonesia’s case nickel, ore exports. Rare earth metals are the exception that proves the rule: the lollipop for that group is led not by China but by Vietnam at 37.5 percent, with Australia at 24.6 percent and China at 21.1 percent, a distribution we return to below.

Each lollipop is a single year’s snapshot of exporter shares within a narrowly defined product set. It says nothing about reserves, refining capacity, or downstream processing — only about who shows up as the exporter of record for these specific HS92 codes in 2024.

Tracking the export-concentration index (a fractional Herfindahl-Hirschman index bounded between 0 and 1, where 1 is a single exporter) over time shows these are not static structures. The cobalt chain concentrated sharply: its index climbed from 0.065 in 2000 to 0.313 in 2024 as the DRC’s share rose from roughly 7.4 percent in 2000 (then fourth, behind Zambia's 11.0 percent) to 54.6 percent. Graphite tells a parallel story of a Chinese takeover: in 2000 the top exporter was the United States; China then passed every rival, lifting the index from 0.098 to 0.271.

Rare earth metals moved the opposite way. China’s grip on this narrow code peaked above 84.1 percent around 2008 (CHN was the top exporter), pushing the index past 0.7; by 2024 the index had fallen to 0.254 as Vietnam and Australia took export share. That decline does not mean rare-earth supply is unconcentrated in any strategic sense — refining and magnet production remain highly concentrated elsewhere — but at the level of this traded metal code, the market genuinely broadened. Lithium drifted up from 0.251 to 0.351, while copper and nickel ores stayed the least concentrated group at 0.131, spread across many mining nations.

The index is computed only over exporters of record for each code set; it is a measure of trade concentration, not of geological or processing concentration.

Concentration on the export side has a mirror image on the import side. Among importers of the critical-minerals group in 2024, China stands alone: it bought about $73.8B of these ores and compounds, more than four times the next-largest importer, Japan at about $14.7B. China’s own supplier base is comparatively diversified (a supplier concentration index of 0.155, with its top source, Chile, at 30.8 percent), which is what a large, strategically managed buyer would be expected to build.

The bubble plot places import value against supplier concentration, and the contrast is the point. Large buyers such as China, Japan, and Korea sit at low-to-moderate supplier concentration — they can spread purchases across Chile, Peru, Indonesia, and Australia. Smaller or more captive buyers sit higher: Mexico, for example, drew about 92.8 percent of its critical-mineral imports from USA (a supplier concentration index of 0.86), the kind of single-source dependence that the export-side averages hide. Supplier shares here are computed from bilateral source-country flows; they describe where a country’s imports originate, not the shipping routes those imports travel.

This is a source-country exposure proxy. It does not observe vessels, transshipment, or the processing steps between mine and cell, and it should not be read as a map of physical supply risk.

Because trade is reported in dollars, a value chart can rise on price alone. Quantity coverage in these groups is strong — at least 99.4 percent of 2024 group value sits on rows with a non-null quantity — so the value figures can be decomposed against physical tonnage. The results vary by material and reshape the growth story.

Lithium compounds grew about 49.6-fold in value but only about 7.7-fold in tonnage; the gap is implied unit value, which rose roughly 6.4-fold. Much of the lithium “boom,” in other words, is price and product mix, not just more material crossing borders. Copper and nickel ores show the same signature: value up about 15.4-fold against quantity up about 4.8-fold, with implied unit value climbing from roughly $381 to about $1,235 per ton — a 3.2-fold price-and-grade effect.

Not every market is a price story. Cobalt’s implied unit value actually fell over the period (a multiple of about 0.8), so its value growth was driven by volume even as prices softened from their peaks; graphite sat in between, with a unit-value multiple near 3.0. The takeaway is the caveat made quantitative: for several of these minerals, a large share of the headline growth reflects the commodity price cycle rather than a proportional increase in physical supply. Unit values are implied (value divided by reported tonnage across heterogeneous grades and forms) and should be read as directional, not as market prices.

Every figure in this article is computed from CEPII BACI bilateral trade data (HS92 product nomenclature, 1995-2024) via committed, re-runnable queries; values are nominal US dollars converted from the source’s thousands-of-dollars unit. The exporter and importer shares, the concentration indices, and the value figures are measured. The grouping of these particular HS92 codes into battery, electronics, and grid-relevant baskets is an editorial choice, which is why the overall evidence status is mixed rather than purely measured.

Three limits matter most. First, raw materials and processed materials carry different HS codes: the lithium, cobalt, graphite, and rare-earth codes here capture specific traded forms, not the full mine-to-battery chain, so the article does not claim to cover refining, cathode and anode manufacturing, or finished cells. Second, the rare-earth picture in particular understates strategic concentration, because most rare-earth value moves embedded in magnets and components under other codes. Third, source-country shares are an origin proxy, not route or vessel data. Read together, these caveats keep the claim narrow: small markets, fast growth, persistent export concentration — measured at the level of specific traded commodities, and no further.