A system-dynamics model of asteroid-mined platinum predicts the PGM price holds near current levels until essentially all supply has shifted off-world, then collapses toward the cost of space extraction

Nachtrieb and Smith (arXiv:2607.06806, "Will AstroForge Collapse the PGM Market?") build "a non-steady system dynamics model of the PGM market" — platinum-group metals — to ask what happens to price when a much cheaper off-world supply becomes available. Naive intuition says unlimited new reserves crash the price immediately. The model predicts the opposite trajectory: the price holds near the current terrestrial level through most of the transition, because as long as terrestrial sources remain in the market the price stays anchored to their higher costs — space suppliers (the paper expects follow-on entrants, not a lone firm) take the margin rather than undercutting. The price falls only at the end, when the last of terrestrial supply is displaced — "the market price for PGM will eventually drop towards the much lower cost of asteroid mining, but only after the entire supply has shifted off-world."

The mechanism is deployment lag: an effectively unlimited reserve gated by a capacity-constrained delivery channel. The gap between the old price and the new (much lower) space-extraction cost floor is not competed away early; it persists as margin for whoever supplies from space first — the abstract's "huge fortunes will be made" — right up to a cliff-like collapse once terrestrial supply is fully displaced. This is a candidate general shape for any unlimited-reserve, capacity-constrained disruption, not only PGM. The company motivating the paper is described in claim-astroforge-mtype-asteroid-pgm-startup.

[unverified-mechanism — needs primary]. This is a model's prediction, quoted from the abstract; the model file and paper body were not examined. Verification routed to question-verify-astroforge-pgm-model-dynamics.

Bridge (2026-07-11): This is the collapse case of the "value = unforgeable cost of production" law. The same law's design case is claim-nakamoto-bitcoin-leaned-on-wei-dai-b-money — Szabo's bit gold (Nakamoto's precursor) was engineered so costliness cannot be forged away, the escape from exactly the cheap-new-supply collapse this model predicts. See 10-inbox/raw/2026-07-11-hop-unforgeable-costliness-bridge.md.

Bridge (2026-07-12): The historical analogy this model's own seed question named — aluminium after Hall-Héroult — turns out not to trace this curve: claim-cheaper-extraction-disruptions-fall-monotonically-not-hold-then-collapse finds both aluminium and AI inference fell monotonically (Wright's law), not hold-then-collapse. That doesn't falsify this model, but it does mean the hold-then-collapse shape needs this note's specific deployment-lag condition, not just "cost of extraction drops a lot."

Bridge (2026-07-18): The commentary's named classical counterpoint, Hotelling's rule, is now in the vault — see claim-hotellings-rule-predicts-shape-distinct-from-cost-of-production-cluster. It confirms this note's shape and Hotelling's are genuinely different predictions (rising vs. hold-then-cliff), not the same mechanism under another name; the deployment-lag-as-scarcity-rent connection flagged in this note's commentary remains an unverified lead, not yet chased.

Bridge (2026-07-23): A second full-text read of the same paper (this time via extract_pdf on the PDF directly, not just the abstract) re-confirms Table III's numbers unchanged and adds three things this note didn't carry: the price-hold is a stated modeling assumption, not a derived equilibrium result (claim-astroforge-pgm-terrestrial-price-anchor-is-assumed-not-derived); the model's actual stock-flow structure and named feedback loops that produce the shape (claim-astroforge-pgm-three-tier-supply-and-named-feedback-loops); and a demand-assumption discrepancy between the paper's simplified table and its full simulation (claim-astroforge-pgm-demand-assumption-differs-table-vs-simulation). No new claim-note was written for the Table III/Figure 4 confirmation itself — it's the same finding this note's 2026-07-09/07-12 audit already recorded verbatim.