Market transmission case study — the 2025–26 copper supply-shock cluster (Grasberg mud rush + Kamoa-Kakula flooding): the first PHYSICAL-disruption trigger
R57 Rung-1 case (2026-07-07, user directive). Every prior case in this corpus is triggered by a state action (export ban, licensing, tariff). This one tests whether the same transmission engine carries a physical supply shock at a named node — a mud rush and a flooded mine — with no government anywhere in the trigger. It does, with one structural inversion (Ch.2 below). Research, not investment advice. All figures traced to operator disclosures and contemporaneous reporting listed at the end; day-move percentages are as-reported intraday/close figures, not beta-adjusted abnormal returns (honesty note at the end).
The trigger cluster
| event | date | what happened | supply effect |
|---|---|---|---|
| Kamoa-Kakula seismic flooding (DR Congo; Ivanhoe Mines 39.6%) | 2025-05-20 | Seismic event + major water inflow flooded the Kakula underground workings; stoping suspended, mine "reset" as Kakula 2.0 | 2025 guidance cut to 370–420 kt (11-Jun-2025); 2026 restored to 380–420 kt (3-Dec-2025); then re-cut −23% to 290–330 kt and 2027 to 380–420 kt (1-Apr-2026) |
| Grasberg mud rush (Indonesia; Freeport / PT-FI) | 2025-09-08 | ~800,000 t of wet material rushed through the Grasberg Block Cave to multiple levels; two workers confirmed dead by 20-Sep, five more missing at the force-majeure date; mining suspended | Force majeure declared 23/24-Sep-2025; Q3 sales guidance cut; 2026 production ≈ −35% vs prior; GBC = 50% of PT-FI P&P reserves and ~70% of forecast output through 2029; cumulative loss ~591 kt through end-2026 |
Together the cluster flipped the 2026 copper balance from a small forecast surplus to a ~400 kt deficit — a multi-year, DATED tightening with no diplomatic off-ramp.
Verdict — what actually traded
| # | expression | move | channel |
|---|---|---|---|
| 1 | Any-domicile producer longs, day 1 (Antofagasta, Teck, Glencore) | ANTO +7.4%, TECK +5%, GLEN +3% on 24-Sep-2025 | every OTHER producer of the material gains — no bloc filter (see Ch.2 inversion) |
| 2 | The affected-asset trap (Freeport) | FCX −10.4% intraday 24-Sep; ~−20% over the episode window | the loser is the single asset, not a domicile class; GBC concentration (70% of PT-FI forecast output) made FCX the highest-conviction short of its own supply shock |
| 3 | The spot expression (LME 3M copper) | +2% to $10,172/t day 1 (15-month high) → $11,000/t 8-Oct (third time ever) → record $12,282/t before Christmas 2025 | physical scarcity with a dated duration re-prices the curve, not just the headline day |
| 4 | The double-hit (Ivanhoe) | IVN cut guidance three times over the cycle; the 1-Apr-2026 −23% re-cut "stunned the market" after the Dec-2025 restoration | a disrupted asset's OWN recovery schedule is the residual risk — see the schedule-slip asymmetry below |
| 5 | The refining-capture premium (Aurubis) | European copper premium raised to a record $315/t | mine-supply scarcity flows to whoever holds ex-disruption processing capacity — the node layer, not just the miners |
The structural inversion vs the policy lane (Ch.2 — read this first)
In every policy case, the winners pass a domicile filter: ex-controller-bloc producers only (a China ban helps non-China supply). A physical disruption has no controller and no bloc — a Grasberg flood helps other Indonesian copper too. The Ch.2 "domicile trap" inverts into an asset trap: the only structural loser is the operator of the affected asset (FCX −10.4%) and co-located / correlated-risk peers. Winners = producers with spare or growth capacity, ANY domicile — which is exactly what printed on 24-Sep (a UK-listed Chilean miner, a Canadian, and a Swiss trader all up together, a combination no policy event produces).
The no-round-trip differentiator — refined, not just confirmed
Our directional event study found policy SUSPENSIONS flip sign 2/2 (China threatens → walks it back → the scarcity premium unwinds). A flooded mine does not un-flood at a negotiating table — the R57 thesis. This cluster confirms it and adds the asymmetry: physical events carry schedule risk instead of reversal risk, and it slips in only one direction. Ivanhoe restored 2026 guidance to 380–420 kt in December, then re-cut it 23% in April; Freeport's "full recovery" migrated from 2027 (company/BMO, Sep-2025) toward 2028 in later street estimates. Recovery timelines lengthen; they do not shorten. For the scarcity trade that asymmetry is a tailwind — the dated duration is a floor, not a midpoint. (The honest converse: the entry premium decays as each recovery milestone is hit — dewatering %, restart phases — so the position has a published, checkable countdown the policy lane never offers.)
Already-priced check + confounders (Ch.3)
- Copper was NOT at rest before the cluster: the US Section 232 copper tariff
(Jul-2025) had already distorted COMEX-LME spreads and pulled metal into the US, and the 24-Sep day-move had a Fed-outlook component (contemporaneous reporting credits "Grasberg disruption, Fed outlook" jointly). Day-1 moves overstate the clean disruption signal.
- The May-2025 Kakula event was the cluster's leading leg: copper's grind
higher through H2-2025 already carried Kamoa scarcity before Grasberg hit. By the time the July-2026 sell-side notes (the trigger for this study) named both mines, LME had printed $12,282 — the cluster is now LARGELY priced at the metal level. The residual expression is single-name (recovery-schedule slippage, premium capture), not the spot long.
The sequel watchlist (Ch.5)
The event class is now defined: single assets whose loss moves the global balance. That is precisely what the node registry ranks (the R58 chokepoint board). The physical-disruption watch = the board's sole-node and low-n entries plus the mega-mines outside it — any operator guidance cut, force-majeure notice, or exchange filing on one of those names is the Rung-2 feed. Same 1k-week-free discipline as the policy lane: the primary source is the operator's own disclosure, so §6 verify-or-don't-file maps unchanged.
Honesty block (quality gate)
- n = 2 events — this is a transmission map, not a cohort statistic; no
claim survives the hedge-fund-grade bar (no beta adjustment, no clustered SEs, no pre-event window test). Day moves are as-reported, not CARs.
- Fatalities: two confirmed by 20-Sep-2025 with five missing at the
force-majeure date; we have not independently verified the final toll.
- We did not verify IVN's share move on the 1-Apr-2026 re-cut (source
paywalled); "stunned the market" is the reporting's characterisation.
- Winners listed are the large liquid names the day-1 tape verified. The
policy-lane lesson (the +100–200% lives in small pure-plays) is NOT yet tested for physical triggers — micro-cap copper developers around the event dates are the open research item for this lane.
Product implication
The transmission engine is trigger-agnostic — confirmed. A physical event produces the same sellable artifact (named winners with channels, the trap, the sequel watchlist) with two upgrades: the catalyst is baked in (no diplomatic round-trip) and the countdown is published (operator recovery milestones). Rung 2 (node disruptions: schema, ex-affected-asset alternatives mode, guidance-cut feed) is now worth ranking.
Sources: FCX investor release "Freeport Provides Update on PT Freeport Indonesia Operations" (force majeure, 800 kt wet material, fatalities, 35%/2026, GBC reserve shares); Ivanhoe Mines news releases 11-Jun-2025, 3-Dec-2025 (380–420 kt 2026, smelter start 1-Dec-2025) and the 1-Apr-2026 guidance release (290–330 kt 2026 / 380–420 kt 2027, Kakula 2.0); Bloomberg 24-Sep-2025 (FCX −10.4%, GLEN +3%, TECK +5%, ANTO +7.4%); MINING.COM (LME $10,172 day-1; $11,000 8-Oct; Ivanhoe re-cut); Benchmark Source LME-week preview ($12,282 record, ~591 kt cumulative loss, ~400 kt 2026 deficit, Aurubis $315/t premium); The Deep Dive 1-Apr-2026 (guidance table). All primary where available; secondary reporting used only for market prints.