Question: Copper supply chain electrification bottleneck Verdict: add-to-watchlist

Chain map — physics first

Price truth: all price / RSI / trend figures below are read from the per-watchlist scan summaries at summaries (macro-commodities; FCX/SCCO live in ai-infrastructure), as-of 2026-06-26 close (~5 trading days stale). Code computes the math; nothing is derived in prose. FCX has no current summary entry — only a stale target-file price ($70.15 / 2026-06-16), flagged as a coverage gap.

Copper's chain is short and old, which is exactly why supply is inelastic. Ore → concentrate → smelt/refine → cathode → semis (wire/rod/tube) → end-use. The binding constraint is upstream (mined concentrate), not downstream fabrication.

Layer 1 — Mined concentrate (the actual bottleneck). Ore grades are in secular decline (many big mines now <0.5% Cu), and a greenfield mine is a 7–15yr permit-to-first-pour project. This is the physics: you cannot flex mined supply on an AI/EV/grid demand timescale.

• FCX Freeport [tracked — ai-infrastructure] — largest US-listed copper name; ~copper-dominant with gold/moly byproduct. Owns its concentrate. Durable large-cap.
• SCCO Southern Copper [tracked — ai-infrastructure] — cleanest pure copper, lowest-cost quartile, huge reserves; but Grupo México control (minority-holder discount) + Peru/Mexico jurisdiction risk.
• TECK Teck [tracked — macro-commodities] — post-steelmaking-coal-divestment now a copper-growth base-metals name (QB2 ramp + zinc). Cleaner copper story than it was.
• ERO Ero Copper [tracked incidentally — monster-discoveries] — Brazil small-cap pure-play, Tucumã ramp; higher beta, spec.
• HBM Hudbay [gap] — mid-cap (Peru Constancia + Manitoba + AZ Copper World); higher beta. Note: ticker HBM = Hudbay, not "high-bandwidth memory".
• Diversified [gap/foreign]: BHP [incidental — monster-discoveries], RIO [gap] — copper is the growth leg but iron ore is still the profit engine → diversified, not clean copper.
• Foreign primaries [gap]: Antofagasta (ANFGF), Ivanhoe (IVPAF / IVN.TO), First Quantum (FQVLF / FM.TO), Lundin, Glencore (GLNCY — also smelting/trading).

Layer 2 — Smelting/refining + the TC/RC signal (the fresh angle). Smelters charge miners a Treatment & Refining Charge (TC/RC) to turn concentrate into cathode. China overbuilt smelter capacity; when mined concentrate is tight relative to that capacity, TC/RC collapses — and 2025–26 spot TC/RC has run near/below zero, an unprecedented tell that concentrate (Layer 1) is the true scarce resource. Directional reads:

• Helps integrated miners that own concentrate (FCX, SCCO, Antofagasta, Ivanhoe, First Quantum) — they capture the tightness.
• Hurts merchant/custom smelters with no captive mine — Aurubis (Germany, NDA.DE / AIAGY [foreign]), Chinese smelters, and the smelting sleeves of diversifieds (Glencore, Sumitomo). This is the avoid/short side, not a long pick.
• The specific current spot TC/RC figure is not on disk — see follow-ups. The mechanism is durable; the live number needs verification before it's cited.

Layer 3 — Secondary / scrap (the elasticity valve). Recycled copper is ~15–35% of refined supply and is the fastest-responding source when price spikes — it's the reason a copper "shortage" caps out as a price-and-substitution problem, not a physical stockout. It caps the tail of the bull case. No clean US-listed scrap pure-play; captured inside diversified recyclers/traders (Glencore, Aurubis, scrap processors) — no investable pure-play [none].

Layer 4 — Expression / ETF.

• COPX [tracked] — miners basket; carries the operating leverage (and holds the Layer-1 concentrate names).
• CPER [gap] — copper-futures ETF; the pure spot-price expression, no operating leverage.

Investable picks

Cleanest US-listed longs (all already tracked, just scattered):

• FCX — the anchor. Diversified-into-copper large-cap, owns concentrate (TC/RC winner), trend-hold quality. Only on-disk price is the target file's $70.15 as of 2026-06-16 (stale, coverage gap — not in any summary I hold; see follow-ups).
• SCCO — highest-margin pure-play but governance/jurisdiction discount. Trend-hold-able as a high-dividend compounder.
• TECK — $58.22, RSI 40.4, weak-down, +13.2% vs SMA200, -10.6% 30D, +20.1% 3M, +44% 1y (macro-commodities summary, as of 2026-06-26). Still well above SMA200 = intact uptrend in a pullback.
• COPX — $76.18, RSI 39.1, weak-down, +2.4% vs SMA200, -11.5% 30D, +6.4% 3M, +74.6% 1y, alpha1y +54.8, -23.8% from 52wk high, regime pullback (as of 2026-06-26).

Higher-beta / spec (not core trend-hold): ERO, HBM.

Cleanest expression of the TC/RC angle: own the integrated concentrate-holders (FCX/SCCO) over CPER — but note COPX already is a basket of those miners, so COPX captures ~most of the "own the tightness winner" trade. CPER (futures) is the one expression that does not capture it. So the TC/RC edge is: prefer COPX/FCX/SCCO over CPER, and avoid merchant-smelter names.

Trend-hold read (do NOT treat as entry): the secular uptrend is intact (COPX +74% 1y, +127% 5y; TECK +44% 1y) but the tape is in a macro correction — COPX weak-down RSI 39, barely above SMA200, -11.5% 30D. This is a strong-dollar-driven metals selloff (UUP RSI 65.9, strong-up breakout, near 52wk high — the ai-power-bottleneck README's 2026-06-26 note explicitly flags the dollar "crushing commodities"). Not a buy-now flag/breakout setup; it's a correction inside an uptrend. Verify current technical state before any entry — the re-entry trigger is a dollar-strength unwind + re-base, not this level.

Reconciliation

vs COPX/FCX (current coverage). Copper demand (AI + EV + grid) is a genuinely consensus macro trade and it's already covered: COPX/TECK in macro-commodities, FCX/SCCO in ai-infrastructure, ERO/BHP incidentally in monster-discoveries. Simply owning COPX/FCX already expresses the demand thesis and already holds the concentrate-rich miners. The only incremental edge this trace surfaces is (a) the TC/RC collapse as a supply-tightness confirmation + a which-name discriminator (integrated miners > merchant smelters > CPER), and (b) that the copper names are scattered across four watchlists with no coherent copper grouping — you can't run a copper-specific scan today. Neither of those clears the bar for a durable new perspective; both are cheaply captured by a watchlist grouping + a monitoring note. The thesis-idea's trigger has also cooled: the source cited COPX "+30% vs SMA200"; it's now +2.4% — the setup that motivated the idea has substantially mean-reverted, lowering urgency.

vs ai-power-bottleneck. This does not belong there. That perspective is about electrons — generation, prime movers, PPA/turbine backlog — priced on contracted backlog and IPP economics. Copper is the metal, priced on LME/concentrate balance and the dollar. Folding copper into ai-power-bottleneck would be a category error; the two share only a demand narrative, not a supply object or a pricing mechanism. Copper is more naturally a macro-commodities resident than an AI-power resident.

Verdict + reasoning

add-to-watchlist (explicitly NOT promote-to-perspective, NOT fold-into-ai-power-bottleneck).

Reasoning: copper is a crowded, well-known consensus trade already covered by COPX/FCX/TECK/SCCO on disk; the demand thesis carries near-zero incremental edge over just owning COPX. The one genuinely fresh, non-consensus element — the 2025–26 TC/RC collapse — is real and useful as a supply-tightness confirmation + name-discriminator, but it's narrow and is mostly expressed by owning the same integrated miners COPX already holds. That doesn't justify the overhead of a dedicated perspective (log, refresh cadence, catalyst table). What IS cheap and additive: group the scattered clean US names into one copper watchlist so the chain becomes scannable, and leave a monitoring note. Defer a real perspective until there's something to actually trade that COPX doesn't already give you — i.e. a TC/RC-driven divergence (long integrated-miner / short merchant-smelter) or a clean idiosyncratic pure-play. Ponytail: a watchlist + a note is the right-sized artifact here; a perspective would be over-building for a consensus macro exposure we already hold.

Proposed mutations / follow-ups

Watchlist (clean US tickers only):

• Add a copper tracking list: FCX SCCO TECK ERO HBM COPX CPER.
  • Creates a coherent copper supply-chain grouping. FCX/SCCO/TECK/ERO/COPX already exist elsewhere (dedupe, not new exposure); HBM (Hudbay) and CPER (copper-futures ETF) are the only genuine coverage-gap adds. AA is aluminum — do not put it in a copper list.

Gated foreign/ADR coverage-decisions (research-only; keep out of watchlist/trade rows unless authorized):

• Antofagasta (ANFGF), Ivanhoe (IVPAF / IVN.TO), First Quantum (FQVLF / FM.TO), Glencore (GLNCY), BHP, RIO — thin/OTC US proxies; add as _securities.json security-family rows only if we decide to track the foreign primaries. Aurubis (NDA.DE / AIAGY) is the TC/RC loser illustration, not a long — track only if we want to build the divergence trade.

Perspective: none now. Revisit a copper / electrification-metals perspective ONLY if (a) TC/RC divergence becomes a tradeable long-miner/short-smelter setup, or (b) a clean idiosyncratic pure-play emerges. Leave breadcrumb: considered, dropped — consensus demand trade already covered by COPX/FCX; TC/RC edge too narrow for a standalone perspective today.

Deep-dive follow-ups (feasibility-gated):

• FCX deep-dive refresh (anchor name; existing target file is stale at $70.15/2026-06-16) — verify current technical state + TC/RC exposure. Feasible via Massive/EDGAR.
• SCCO deep-dive (governance discount + cost-curve position) — feasible.
• Skip any "10y multiple-percentile" style asks — spot .info can't support it.

Monitoring note (not a task): copper re-entry trigger = UUP (dollar) strength unwind + COPX re-base above SMA20; supply-tightness confirmation = spot TC/RC staying near/below zero. Both are watch-items, not buy signals at current levels.

Sources

Price, RSI and trend figures read from the desk's validated daily scan summaries (summaries). No number in this note was computed in prose.