Start with the tape. LME copper settled around $13,662.50/t on 15 June 2026, sitting just under the ~$13,800 it brushed earlier this month (MacroMicro; MoneyMagpie). That follows a genuinely violent January: copper printed an all-time intraday $14,527.50/t on 29 January, the biggest single-day jump since 2008, after first clearing $12,000 only in December 2025 (Benchmark).
This is not a vibe. It's a supply accident meeting a demand story.
When the world's second-biggest copper mine loses ~591kt and the annual smelter charge falls to literally zero, the price isn't speculating — it's reporting.
The supply latency is physical, not theoretical. Grasberg's September 2025 mud rush took the world's second-largest copper mine offline; Benchmark pegs the loss at roughly 591,000 tonnes between September 2025 and end-2026, on top of a fatal tunnel collapse at Codelco's El Teniente and flooding at Kamoa-Kakula ($IVN) (MINING.COM). The clearest tell sits midstream: Antofagasta ($ANTO) settled the 2026 TC/RC benchmark at $0/t and $0/lb, the lowest ever and down from $21.25/$2.125 in 2025 — miners no longer pay smelters to turn concentrate into metal (MINING.COM; IEA).
Inventories and term structure confirm the squeeze. On-warrant LME copper fell ~22% to 89,725 tonnes in mid-January, a six-month low, as metal flowed to the US ahead of a pending refined-copper tariff decision — widening the COMEX-LME premium (Business Recorder). The demand pull underneath is grid, electrification and AI data centres; that's the structural leg, not the catalyst.
Now the companies, held to the cost curve.
Freeport-McMoRan ($FCX) (FCX) is the high-realisation, broken-asset trade. Q1 2026 was strong on paper — revenue $6.2bn, adjusted EBITDA $2.5bn, EPS $0.61 versus $0.24, realised copper $5.78/lb — yet the stock fell ~12% to $61.67 when management said Grasberg's Block Cave will run at only ~60% of capacity through mid-2027 (Investing.com). In our view that's the right reaction: price leverage is wonderful until the tonnes aren't there.
Ivanhoe Mines (IVN) is the volume-restart story. It confirmed 2026 guidance of 380,000–420,000t copper and 500,000–540,000t for 2027, with its new Kamoa-Kakula smelter producing 99.7% anode at ~500t/day; 2026 sales should exceed production as ~20kt of stranded concentrate is destocked (Ivanhoe). Friedland's pitch is vertical integration in the DRC — we like the trajectory and remain watchful on underground rehab after the flood.
Antofagasta (ANTO) is the cost-curve winner. Q1 output dipped 8% YoY to 143,000t on planned lower grades, but net cash costs fell ~30% to $1.08/lb, Centinela runs near 34c/lb, and management held full-year guidance at 650,000–700,000t at $1.15–$1.35/lb; CEO Iván Arriagada said the "copper price remains constructive in 2026" (bez-kabli). That is what you want at the top of the cycle.
BHP ($BHP) (BHP) is the boring, bottom-of-cost-curve cash machine. Escondida hit record concentrator throughput, FY26 group copper guidance rose to 1,900–2,000kt, and HY26 copper EBITDA jumped 59% to a record $8.0bn — copper's first time as the majority of group earnings (BHP 6-K).
The risk: somebody's offside. StoneX flagged LME net longs near the 80th percentile while SHFE specs sit net short — opposing views that historically resolve fast (S&P Global). Own the low-cost pounds. Rent the leverage.