Robotics & humanoids — the physical-AI materials call (new) Positive — own the bill of materials, not the robot makers ▲
Sources: McCracken · astrid-wilde · Updated: 2026-SEP-11
The bull case for robotics that does not require picking a robot manufacturer. McCracken (AI/mathematics PhD, 2026-AUG-28, Value Hive) argues the information era's last rung is being climbed by brute force — digital compute is general but inefficient, so scaling means more GPUs, more memory and more power — and the next efficiency step is purpose-built analog hardware, which cannot be built at scale without an era of general robotics ("not necessarily humanoid — could be tiny little things that help you position a cable"). Robots are therefore a materials demand shock before they are an equity theme. Sizing method: take an operator's stated unit ambition, multiply by per-unit material intensity, divide by annual mine supply — "every Optimus Tesla robot needs an ounce of silver… when Elon says I want to put one of these in every household on the planet, you can very quickly be like, wait a second, how many ounces is that?" Silver first, then aluminium for mass. The equity leg is explicitly not buyable yet: he passed on the Unitree IPO on price alone, "nothing's screaming cheap," and is waiting for a moment "basically like buying NVIDIA in 2022 — assuming that they can get the commodities." End state he wants: "some mix of robotics with miners — own the people selling the robotics companies the materials, and on top of that the robotics companies themselves." Secondary risk raised: Chinese-made robots in Western homes are a policy exposure as much as an investment one — "or they just turn off and you lose a pile of money." Astrid Wilde (2026-SEP-11): the future is non-humanoid, single-task robots (Matic, the dishwasher), trained on large volumes of human task data - 'demand for data continues to be underestimated'. She declines to build a public physical-AI portfolio: Fanuc and ABB are the only real public robot makers, everything else is a derivative bet, and 'all of the most interesting stuff happening in robots is still in private markets' (conflict stated: IPO-stage robotics names are her customers).