Q3 Concentration: Midstream Median Top-1 Hits 75% — Avg Refine 72%, Gallium 99%
Q3 concentration lens on chokepoint commodities after IEA GCMO 2026: median Top-1 73.1% across 16 stages, 10 stages ≥70% Top-1, midstream median 75%, avg refining (ex-REE) 72% (+2 pp), while rare-earth refining eases to 85% and copper smelting capacity sits near 50% China.
Loading interactive charts…
Our 2026 concentration print answered the first distribution question on USGS Mineral Commodity Summaries 2025 / secondary midstream carries: median Top-1 65% across twenty stages, 8 stages above 70% Top-1, China in 12 Top-1 seats, gallium refined near 98%. The Q3 midstream update then restated the tip with IEA Global Critical Minerals Outlook 2026 — average top refining-country share (ex-REE) at 72%, rare-earth refining eased to 85%, lithium chemicals ~70%, China copper smelting capacity near 50%, and the 2026 copper TC/RC settle at $0/t. This Q3 concentration lens asks the desk follow-up those vintages imply but do not score as a single distribution: how concentrated is the system at the top now — Top-1, Top-3, and HHI — once the newest midstream print replaces the prior carry?
The interactive dashboard above is built as that lens. Toggle Vintage delta, Ranked shares, Mine→plant, and Risk & HHI. Filter by stage and Δ direction. The punchline is deliberately three-sided. On the midstream tip, median Top-1 across processing / smelter / recycle rows prints about 75%, against a mine-stage median near 69%. On the IEA headline meter, average top refining-country share (ex-REE) sits at 72% — +2 pp versus the 2023 baseline the Outlook carries. On the extreme, gallium refined is still near 99% China, while rare-earth refining is the rare durable ease (90% → 85%). China still holds 11 of 16 Top-1 seats in this Q3 table.
The Q3 concentration scoreboard
| Lens | Top-1 | Thick top | What moved vs prior concentration |
|---|---|---|---|
| Median across 16 stages | 73.1% | Top-3 median 88% | Prior median Top-1 was 65% on a 20-stage MCS/secondary ledger |
| Stages with Top-1 ≥ 70% | 10 / 16 | — | Prior print: 8 / 20 |
| Midstream+ median Top-1 | 75% | Mine median ~69% | Processing still hotter than the pit |
| Avg refine (ex-REE, IEA) | 72% | — | +2 pp vs 2023 IEA baseline |
| REE refining / separation | 85% | Top-3 ~96% | −5 pp vs prior ~90% carry |
| Lithium chemicals | 70% | Top-3 ~88% | +5 pp vs prior secondary ~65% |
| Graphite anode | 90% | Top-3 ~95% | Flat share; disruption risk ~$300B/yr |
| Copper smelting capacity | 50% | — | 2005→2025 path; TC/RC settle $0/t |
| Gallium refined | 99% | Top-3 ~99.5% | Still the monopoly extreme |
| China Top-1 seats | 11 / 16 | 69% of leaders | Prior: 12 / 20 |
Read the table as a family of market shares, not one number. Median Top-1 tells you the typical stage tip. The ≥70% count tells you how often the tip is extreme. Midstream vs mine medians tell you whether the binding chokepoint is the pit or the plant. The IEA average-refine meter is a different object again — a cross-mineral refining average the Outlook publishes as a security headline. Analysts who quote only copper mine shares understate smelter capacity; analysts who quote only rare-earth ease understate graphite anode and battery-recovery monopolies that did not ease.
Vintage delta: what tightened, what eased
Filter the dashboard to Vintage delta. The grouped bars rank the largest absolute percentage-point moves from the prior concentration carry to this Q3 restatement. Copper smelting capacity’s +35 pp path (15% → 50% China, 2005→2025) dominates the chart because it is a capacity geography restatement the prior Midstream secondary table never fully owned. Nickel refined-supply growth capture (+25 pp to ~75%) and battery material-recovery capacity (+20 pp to ~90%) follow — both IEA prints that say secondary supply and incremental refine growth are concentrating even when some stock shares look flat.
The easers matter because they are rare. Rare-earth refining / separation prints 85% Top-1 (−5 pp), with a project path toward ~70% by 2035 if US and Malaysia capacity delivers. Natural graphite mine eases slightly on the MCS 2026 hold (79.4% → 77.8%). Cobalt refine is roughly flat near 75% — and the Q3 update is explicit that DRC export-quota risk is a volume story, not share relief. Graphite anode stays ≥90%: the share did not move; the downstream-risk framing did (~$300B/yr outside China on a full trade disruption path).
Across the sixteen-row table, 7 stages tighten, 2 ease, and 7 stay flat. That mix is the Q3 contribution: the tip did not uniformly soften just because one magnet midstream meter improved.
Ranked shares: Top-1, Top-3, and HHI
Switch to Ranked shares. Toggle Top-1, Top-3, or HHI, then filter stage and sector. On Top-1, the order is familiar but restated: gallium ~99%, graphite anode and battery recovery ~90%, REE refine 85%, then the mid-70s cluster (graphite mine, cobalt mine/refine, nickel growth capture, average refine). Lithium chemicals now clear the 70% line. Copper refine and smelt sit near 50% — not a monopoly, but a decisive midstream tip next to Chile’s ~23% mine lead.
Top-3 bars make the “thin remainder” visible. Eleven of sixteen stages clear ≥85% Top-3. Even where Top-1 is “only” the high sixties or seventies, three capitals often clear almost the entire world. Analytical country-share HHI (0–10,000) puts thirteen stages at or above 2,500, with a median near 5,525 — a highly concentrated band on the conventional antitrust-style scale, even though these are country shares, not firm HHIs, and must be read as derived meters.
The producer seat chart answers a different question: who leads. China still occupies 11 Top-1 seats (69% of leaders in this table). Indonesia leads the nickel growth-capture row; DRC leads cobalt mine; Australia leads lithium mine; Chile leads copper mine; Morocco leads phosphate fertilizer exports. Plural leaders at the pit do not imply plural midstream.
Mine → plant: where leaders flip
Open Mine→plant. The slope panel is the geographic punchline copper and lithium desks already trade intuitively: Chile ~23% of mine copper versus China ~50% of smelting capacity / refine; Australia ~37% of lithium mine versus China ~70% of lithium chemicals; DRC ~74% of cobalt mine versus China ~75% of refine; China already leads graphite mine (~78%) and still tightens further into anode processing (~90%); rare earths stay China-led from mine (~69%) into separation (85%, eased but still extreme).
Beside the slopes, the copper smelter stress meters bind capacity share to fees. China utilisation near 85% versus ~70% outside China, and a $0/t 2026 TC/RC annual settle, answer why “custom smelters will always be there” is no longer a free assumption. Pair that with the August Pink Sheet / spot TC follow-through if you need the price-tape vintage; this post keeps the share meter.
For single-metal geography deep dives keep copper mine vs refinery and natural graphite mine concentration open beside this lens.
Risk, HHI bands, and the investment paradox
Risk & HHI folds three objects desks often keep in separate tabs. The HHI donut shows how many filtered stages sit in extreme (≥5,000), high (2,500–4,999), moderate, and plural bands. The reliance × Top-1 scatter puts US net-import reliance on the x-axis and Top-1 share on the y-axis; bubble size tracks substitution difficulty. Gallium, graphite anode, and several battery midstream rows sit upper-right: high reliance and high Top-1. Lithium chemicals sits high on Top-1 with lower US reliance — a reminder that concentration and import dependence are related but not identical meters.
Downstream disruption risk bars put IEA dollar framing next to Top-1: graphite anode disruption near $300B/yr outside China; rare-earth full-chain disruption framed near $6.5T. Those are scenario envelopes, not base-case losses — but they translate share monopolies into balance-sheet language.
The investment panel is the paradox: critical-minerals investment −9% YoY, battery-metals capex −20%, lithium capex −40%, while copper capex lifts ~+8% and public-finance support prints near $65B (~4× vs 2023). Private capital is pulling back from some of the hottest midstream tips just as public money tries to seed alternatives — a timing mismatch that keeps concentration sticky even when policy headlines sound diversified.
Caveats and how to read the meters
- This is a Q3 concentration vintage, not a second twenty-stage encyclopedia. For the original MCS 2025 / secondary Midstream ladder use the 2026 concentration print. For share deltas without the Top-k / HHI frame use the Q3 update. For the full research scoreboard use the research ledger.
- IEA and USGS definitions differ. Cross-agency Δ is directional. Lithium chemicals ~70% is an IEA processing narrative against a prior secondary carry; copper refine “just under 50%” sits beside MCS ~48%treat them as consistent direction, not identical accounting.
- Top-3 and HHI are derived. Where agencies publish only Top-1 (or a short country list), Top-3 and HHI reuse prior theme ladders and rounded residuals. Do not quote HHI as a disclosed antitrust filing.
- Mine holds are MCS 2026 anchors where IEA did not restate pits (graphite, cobalt, copper, lithium mine). A flat mine share can still hide a tighter midstream or a revised volume-risk outlook (DRC cobalt quotas).
- Country share ≠ firm share. A 50% China smelting-capacity print can still be many plants; a 90% anode print can still be a short list of processors. Both are chokepoints; they are different legal and contractual objects.
- Disruption dollars are envelopes. The $300B graphite and $6.5T REE figures are IEA-style full-disruption framings, not expected annual losses.
What the tip implies for desks
Physical-input assumptions in energy transition, semis, and fertilizer models often treat midstream capacity as elastic once the mine exists. The Q3 concentration print says the opposite for the binding stages: median midstream Top-1 near 75%, 10 of 16 stages above 70% Top-1, average refining (ex-REE) at 72% and rising, graphite anode still ≥90%, battery recovery now ~90%, and copper’s fee cover at $0/t while China capacity sits near half the world. Rare-earth refining’s ease to 85% is real — and still leaves Top-3 near 96%. Diversification announcements matter; they have not yet moved most of the tip.
Related reading: 2026 concentration print · Q3 midstream update · MCS 2026 update · August price/stress vintage · research ledger · copper geography.