Industry Notes / Global mining: the metal map / Chapter 1 of 12 · Thesis cards

Eleven conclusions

Not a broad mining bull—extreme dispersion

1. 1

   Two extremes in one year

   +18%

   Copper YTD (plus squeeze spikes)

   vs

   −22%

   Gold drawdown from the January peak

   Precious-metal crash and industrial squeeze together are the 2026 mining frame.

   Do not price every metal with one “commodities bull” label.

2. 2

   Copper is now the majors’ top profit source

   54%

   BHP FY26 copper share of EBITDA

   vs

   >50%

   Rio copper+aluminium+lithium profit share

   A decade of capital allocation, not a quarterly fluke.

   Split the commodity mix before talking cycle.

3. 3

   AI pull must be tiered

   Rough intensity: copper ≈ uranium > GOES > silver > aluminium > tin ≫ iron ore.

   BHP: ~$200 bn/year of extra data-center spend maps to about one new 150 ktpa copper mine.

4. 4

   Uranium and acid are underwatched

   94–97

   Uranium term price ($/lb, 18-year high)

   vs

   +160%

   Sulfur YTD move, order of magnitude

   Term over spot means buyers pay for certainty; acid can choke SX-EW copper and uranium mills.

   A second AI line plus a cross-metal bottleneck.

5. 5

   Government takes about 43% first

   BHP’s take including royalties is about 42.9%. Equity gains need a ~0.57 multiplier.

   Elasticity tables that ignore tax overstate shareholder returns.

Hold five first: dispersion, copper over iron, AI tiers, iron misallocation, uranium and acid. Later chapters expand one each.

Source：全球矿业深度调研（一手业绩与 LME / 现货公开价，截至 2026-08-18）

Research notes, not investment advice.