I know what you want: a US-listed, pure humanoid, multi-bagger-upside name. Like buying NVIDIA early — one bet on the sector leader.

I have to say it plainly: that name is basically an empty set on US markets. The sooner you accept it, the less you’ll size heavy in the wrong place.

The hardest chokepoints — all un-buyable

Start with “pure-play leaders.” The hardest, most-moated links in embodied AI — dexterous hands, tactile sensing, the robot brain (VLA foundation models), high-end harmonic reducers — who leads them?

Painful answer: almost all private, listed in China/Japan, or embedded inside big companies. The pure-play robot-brain leaders — Physical Intelligence, Skild — are private; you can’t buy them. The dexterous-hand and tactile leaders are mostly private or Chinese. High-end reducers are Japanese oligopolists. The “real chokepoints” buyable on US markets, you can count on one hand.

So when a US name spikes on a “chokepoint concept,” be suspicious: if the real leaders aren’t here, why is this one a leader? Usually you’re buying “adjacent,” not the chokepoint.

Want upside? What you get is a “pricey + cash-burning” option

Now “big upside.” The high-torque pure-plays buyable on US markets rarely escape two words: expensive, and burning cash.

• Tesla: the only body upside, but diluted by a giant — Optimus is ~14% of valuation.
• Aurora (autonomous freight): ~$12.8B market cap against ~$15M revenue — pure option pricing, burning cash.
• Pony.ai (robotaxi): the most real commercialization evidence — city-level unit-economics breakeven already reached in Guangzhou and Shenzhen — but still a high-risk China ADR.

Common thread: the upside you want costs you a pricey, mostly-loss-making option. No free lunch. One-line tiering: Tesla is steady but diluted; Aurora is purest but burns most; Pony has the hardest evidence but eats an ADR/regulatory discount. The purer the upside, the pricier the risk.

So stop fooling yourself — manage them as options

Accepting the “empty set” actually clears your head. Since US markets won’t give you “pure + cheap + big upside,” stop pretending you can have all three — and build and manage these high-torque names as options:

• Fixed budget: the whole option layer is a small “affordable-to-lose” sum — a total zero wouldn’t shake the portfolio.
• Per-name cap: hard ceiling on each name, never broken because you like it.
• Buy evidence, not narrative: add only on a drawdown plus a de-risking signal (an L3 signal, unit-economics breakeven) — not on a headline.
• Cut the thesis, not the price: when the thesis itself is wounded (tech falsified, cash running out, regulation turns), trim regardless of price.
• Never average down into a core: the moment you do, you’ve secretly turned “a small affordable option” into “a core bet you never intended” — the most common way an option position dies.

You want US-market pure-play upside — the price is a pricey, cash-burning option. But put it in a disciplined basket, and you turn “a gamble” into “a disciplined asymmetric bet.” The real leaders aren’t buyable? Then wait patiently for their IPOs — don’t grab an adjacent stand-in and pretend you’re on board.

Ever hunted for “pure humanoid upside” on US markets? Did you end up with a real chokepoint, or something adjacent? Share below.

(Independent industry and educational research, not investment advice. Companies named are illustrative examples, not recommendations. Data is from public sources and may change. The author and affiliates may hold positions in securities mentioned.)

— Adapted from Embodied Intelligence Investing, Deep Dives XV & XVII

#EmbodiedAI #Humanoids #OptionsThinking