S07.1 · Materials, Chemicals & Mining
Copper, iron ore and nickel extraction, sized $800B-$1.95T depending on scope, where autonomous haulage and AI ore modeling are cutting labor costs and geological-expertise moats.
Diversified and base metals mining is the extraction of copper, iron ore, nickel and related industrial metals. Sizing depends entirely on what gets counted: roughly $800B if the measure is the top 40 global miners' combined revenue (PwC's Mine report), or a $1.95T trajectory by 2035 on an ore-to-metal basis (Precedence Research, 2025) — the spread is a scope question, ore versus processed metal versus producer revenue, not a disagreement about the business. Growth runs 3-5% a year with GDP and Chinese construction activity, which drives roughly 55% of base-metals demand. None of that is where the segment is being re-sorted. The re-sorting is happening on the cost line, where autonomous haulage and AI-driven ore-body modeling are attacking the largest controllable expense and, with it, the value of in-house geological expertise.
Headquarters sit in Australia, the UK/Switzerland (Glencore, Anglo American) and Brazil (Vale); the ore sits elsewhere — Chile and Peru for copper, Australia and Brazil for iron ore, Indonesia and the Democratic Republic of Congo for nickel and cobalt. The top 10 diversified miners take roughly 35-40% of mined value, with thousands of junior explorers underneath. The chain runs from upstream exploration and junior mining through smelting, refining and trading into steel, auto and construction end markets.
These are price-takers with heavy balance sheets. EBITDA margins of 25-40% at mid-cycle look comfortable until set against the capital required to earn them: $1-5B per mine and 10-15 years to develop one. High fixed investment, long lead times and no pricing power over the commodity make the cyclicality structural — no operator executes its way out of it. The common responses are portfolio-level: holding diversified base-metals positions alongside critical-minerals exposure as a hedge against the cycle, or pushing into downstream refining and smelting to capture margin beyond pure extraction.
Labor per tonne historically runs 20-30% of operating expenditure, and it is the line autonomous haulage fleets and AI-based ore-body modeling are collapsing — not prospectively, but at commercial scale today, with autonomous truck fleets running in the Pilbara and in Chilean copper operations. Exploration compounds the effect. AI-assisted ore-body modeling and target generation are shortening discovery cycles and cutting the dependence on in-house geological teams to identify viable deposits. The moat that erodes is the one built on proprietary subsurface data and the most experienced geology staff: smaller and newer entrants can now reach comparable target-generation quality with AI tools instead of decades of accumulated field knowledge.
What that does is relocate advantage inside the operation. It sits less in headcount and tenured geological judgment, more in the scale and quality of operational data a producer can feed its autonomous systems — and in the capital available to run autonomous fleets across a mine's full operating life. A producer on older, non-instrumented assets faces a widening cost gap against peers that have automated haulage and integrated AI exploration tools, wherever the commodity price happens to sit in the cycle.