S07.2 · Materials, Chemicals & Mining
A $280-340B gold, silver and PGM mining market where AI-driven exploration targeting is starting to reverse a decade of rising per-ounce discovery costs.
Precious metals mining — gold, silver and the platinum-group metals (PGM) — runs roughly $280-340B: gold alone is $220-280B (2024-25, Fortune Business Insights/Grand View Research), with silver and PGM adding a further $40-60B. Read the 3-6% annual growth carefully, because it is almost entirely price, not volume. Gold output has held near 3,600-3,700 tonnes annually (World Gold Council) while the price moved above $3,300/oz in 2025 — the industry is not producing more; it is being marked up. AI's clearest near-term commercial effect sits in exploration, where AI-driven targeting is starting to work against a discovery cost that has risen roughly tenfold per ounce over the past two decades.
China, Australia, Russia, Canada and the United States lead gold output; South Africa and Russia together supply more than 70% of the world's platinum-group metals. Concentration differs by metal: the top 10 gold miners produce roughly 30% of global gold, and PGM supply is tighter still. The chain starts with junior explorers and royalty/streaming financiers — Franco-Nevada and Wheaton Precious Metals among them, which fund mine development in exchange for a share of future output rather than operating mines themselves — and runs through refiners and bullion banks into end demand split across jewelry, investment and industrial use.
The economics are a leverage trade on the metal price. All-in sustaining costs (AISC) of roughly $1,300-1,500/oz against a spot price above $3,300 produce current margins above 55%. The same cost structure works in reverse — those margins compress sharply in a price downturn — and it sits on top of heavy capital intensity and emerging-market political-risk exposure that no price rally removes.
The exploration problem is a cost problem. Discovery cost per ounce is up roughly tenfold over the past twenty years as easier-to-find deposits are depleted, and AI-driven exploration targeting combined with satellite data is starting to reverse the trend by making credible target generation cheaper and more accessible. The producer whose advantage was the largest proprietary geological database loses relative position here: as AI and satellite tools democratize target generation, the edge migrates away from incumbents' accumulated subsurface data toward whoever applies the best modeling tools fastest. Commercial-scale impact is a 2-5 year proposition.
Operating cost is moving on a separate track. Autonomous haulage, drilling and processing optimization are already in use at scale in precious-metals operations — the same trajectory base-metals mining ran — and they compress operating cost per ounce independent of anything happening in exploration. Put the two together and the operators best positioned to hold margin through the next price downturn are those that have automated the cost base and adopted AI-driven exploration, not those running on legacy reserve bases and manual target generation.