How demand and the squeeze are estimated
For each material: principal end-use sectors, the clean-energy share of demand, and a demand-growth multiple to ~2040 (demand in 2040 ÷ today) grounded in the IEA Global Critical Minerals Outlook 2024 (Announced-Pledges scenario) and USGS end-use data. The squeeze index = normalised demand growth × companionality — high only when demand is surging and supply is by-product-locked and cannot scale.
Scenario-dependence, handled explicitly. A forward demand multiple is scenario-dependent by construction, so quoting one number hides the real uncertainty. For the six minerals where the IEA publishes a total-demand series, we no longer curate a point estimate — we compute g in all three scenarios from the IEA Critical Minerals Dataset (CC BY 4.0): STEPS (stated policies) → APS (announced pledges, our central) → NZE (net zero), and carry the whole band into the squeeze. The remaining materials keep a curated literature estimate, labelled as such in the table — the IEA’s 37-mineral sheet covers clean-tech demand only, which is a different quantity from total demand and would overstate g. Inputs: IEA Critical Minerals Dataset + USGS × companionality.json → demand.json.
Demand vs supply elasticity — two very different problems
Right = faster demand growth to 2040. Up = more by-product-locked (supply can’t scale). Top-right is the structural squeeze; bottom-right is demand pressure you can still answer with mines.
Every material — demand pull and the squeeze
| Material | pulled by | clean-energy % | demand ×2040 | outlook | by-prod % | squeeze |
|---|
What is actually doing the pulling?
A multiple says how much more; it doesn’t say what wants it. The IEA breaks demand down by end-use technology, so we can decompose the growth rather than assert it: the biggest use by 2040, and the technology adding the most absolute demand between 2024 and 2040 — the growth driver.
| Mineral | demand mix in 2040 | clean-tech share | growth driver 2024→2040 |
|---|
What this opens
The demand arm turns the atlas from a snapshot of where supply sits into a map of where pressure is heading — and, joined to the supply-structure work, separates the materials that need capital and time (lithium, graphite: mine more) from those that need a different playbook entirely (gallium, germanium, rare earths: recovery yield, stockpiles, substitution, because more mines aren’t on the menu). From here the branches are concrete: demand by technology scenario (what a faster EV path does to each squeeze), demand by country/bloc (whose industrial policy pulls which metal), and coupling demand growth to the price series to test whether the squeeze is already showing up in unit values.