The materials with no way out
10 of the 32 materials are at once hard to substitute, essentially un-recycled (<5% end-of-life), and supply-concentrated — so a disruption bites (no substitute), has no secondary supply to fall back on (no recycling), and one country controls the flow. For 8 of the 10, that chokepoint is China's refining. These — not the loudest headlines — are the materials where a shock is both most damaging and least mitigable.
| Material | risk | top exporter | refined in |
|---|---|---|---|
| Beryllium, unwrought | 74 | 🇰🇿 Kazakhstan 89% | 🇺🇸 United States 65% |
| Lithium carbonate | 64 | 🇨🇱 Chile 75% | 🇨🇳 China 74% |
| Phosphorus | 57 | 🇻🇳 Vietnam 49% | 🇨🇳 China 78% |
| Gallium | 57 | 🇨🇳 China 29% | 🇨🇳 China 98% |
| Rare-earth permanent magnets | 52 | 🇨🇳 China 64% | 🇨🇳 China 84% |
| Natural graphite | 51 | 🇨🇳 China 36% | 🇨🇳 China 94% |
| Fluorspar, >97% CaF2 | 51 | 🇲🇽 Mexico 38% | 🇨🇳 China 65% |
| Silicon, < 99.99% | 47 | 🇨🇳 China 40% | 🇨🇳 China 79% |
| Germanium | 44 | 🇨🇳 China 29% | 🇨🇳 China 81% |
| Coking coal | 24 | 🇦🇺 Australia 41% | — |
Concentration is intensifying
With the reconciled trade series now reaching back to 2002, the concentration story is testable, not just visible. 10 of the 32 materials show a rising export-concentration trend significant on an FDR-corrected Mann–Kendall screen (23 annual points; exploratory), and their structural breaks cluster in 2012–2016 (9 of them). The trade-side origin gap runs wide (~15–16pp, with a step at the 2017 HS-vintage join), and China's average network throughput — its centrality as importer-and-hub, not just exporter — rose from 10% to 18%. China's export share of rare-earth magnets climbed 19→64%, of tungsten 9→62%. See it move →
Highest supply risk
- Beryllium, unwrought 74 · un-recyclable
- Ferro-niobium 64 · un-recyclable
- Lithium carbonate 64 · un-recyclable
- Magnesium, unwrought 60 · 13% recycled
- Gallium 57 · un-recyclable
- Strontium carbonate 57 · un-recyclable
The refiner illusion — biggest origin gaps
- Beryllium, unwrought · 🇰🇿 Kazakhstan exports, mined in 🇺🇸 United States +89pp
- Strontium carbonate · 🇩🇪 Germany exports, mined in 🇮🇷 Iran +63pp
- Lithium carbonate · 🇨🇱 Chile exports, mined in 🇦🇺 Australia +51pp
- Aluminium ores / bauxite · 🇬🇳 Guinea exports, mined in 🇦🇺 Australia +48pp
- Phosphorus · 🇻🇳 Vietnam exports, mined in 🇨🇳 China +48pp
- Fluorspar, >97% CaF2 · 🇲🇽 Mexico exports, mined in 🇨🇳 China +27pp
No recycling cushion
High-risk and barely recycled — a disruption has no secondary supply to fall back on.
- Beryllium, unwrought · risk 74, 0% recycled
- Ferro-niobium · risk 64, 0% recycled
- Lithium carbonate · risk 64, 0% recycled
- Gallium · risk 57, 0% recycled
- Strontium carbonate · risk 57, 0% recycled
- Phosphorus · risk 57, 0% recycled
Most-exposed economies
- 🇲🇰 Macedonia [FYROM] 31/100 · $1.1B
- 🇲🇽 Mexico 25/100 · $2.7B
- 🇧🇪 Belgium 19/100 · $2.3B
- 🇨🇦 Canada 19/100 · $2.0B
- 🇯🇵 Japan 18/100 · $30.8B
- 🇺🇸 United States 18/100 · $17.9B
So what do you do about it?
Everything above maps the problem — where supply concentrates and who is exposed. The decision layer turns it into action: who could break each chokepoint, whether the world could absorb a cut, and how exposed any one country really is. Same public trade matrix, three tools.
chokepoints that are a capability, not an ore endowment — no export ban touches them. Who could realistically break each, and who is building alternatives.
Open the decision layer →Cut a chokepoint and reallocate optimally: 4 of 31 materials can’t be covered even if every other exporter doubles; for 8, removal makes concentration worse. With a before/after map.
Run the stress test →An importer × material heatmap: how much of each country’s supply runs through a single source, and how single-sourced it is. The East-Asian manufacturing belt is the most locked-in.
Find your exposure →Computed by build_insights.py from the public data (data.json, flows_2024.json, risk.json). Trade year 2024. Decision-layer figures from breakout.json, ot.json, leverage.json.