Critical Materials Atlas
Method · systemic · host dependency

When the host stumbles

A by-product metal is hostage to the commodity it rides on. So the real supply shock for gallium isn’t a gallium event — it’s an aluminium event. This layer — child of the hostage-metals page — inverts the map to expose the ordinary bulk commodities that silently gate critical supply, and lets you shock any one of them.

How the host-shock is modelled

For every by-product material we split its by-product share evenly across its named hosts: exposure(companion→host) = companionality/100 ÷ number of hosts. Inverting gives, for each host commodity, the critical materials riding on it. A shock of S% to a host cuts each rider’s supply by S% × exposure. A host’s payload weights those exposures by each companion’s supply-risk score — how much critical supply hangs on that one commodity.

Caveat — and why static is the right tool here: even host-splitting is a simplification (real shares vary by deposit and are often undocumented), and it assumes a linear pass-through from host output to companion recovery. Read each figure as an exposure bound, not a forecast. This static first-order design is deliberate, not a shortcut — it is the same logic as the official USGS/EU criticality methods, and a four-way method comparison found it the honest spine for an open-data atlas (a VAR would measure price co-movement, not physical loss; an input–output model can’t even resolve by-product metals; structural supply curves need proprietary cost data). The higher-order ripple is handled separately by the cascade layer. Inputs: companionality.json × risk.json → host_shock.json.

Shock a host, watch the companions bleed

Bars show the central exposure bound; the faint marker is the low–high range from a ±15pp uncertainty on companionality. These are bounds on supply at risk, not forecasts — recovery lags, stockpiles and spot trade absorb part of any real shock.

The bulk commodities that gate critical supply

Ranked by payload = critical supply riding on each host (exposure × each companion’s risk score). These are the single points a systemic supply map should watch.

Host commoditycritical riderswhich materials (by exposure)payload

What this spawns

The host map turns an abstract list of “critical” metals into a small set of chokepoint commodities — aluminium, zinc, copper, nickel, natural gas — whose ordinary market cycles propagate into the critical layer. The next children write themselves: a recovery-yield / recycling lens (secondary supply is the only source that doesn’t depend on the host), and a coupling of this map to real shock scenarios and the price series — do historic aluminium slumps actually show up as gallium squeezes?