Robotics & Physical AI: What's Out There?

Executive summary — free edition

Finnacl Research · September 2026 · The full report covers 12 core companies, 20+ more across the incumbents, autonomy operators, and suppliers, the private humanoid roster, nine robotics funds, and a quarterly-scored early-warning scorecard.

The answer

Three businesses at three maturities — and the market prices them as one.

1. Deployed and profitable. 4.66 million industrial robots are at work in the world's factories; half a million more were installed last year, for the fourth year running. Surgical robots did 3.15 million procedures in 2025 (+18%). Warehouse automation is booked years ahead. AI is making this installed base more valuable — it removes the programming that made automation expensive to change. This is where the money is made today, and it trades at ordinary industrial multiples.
2. Deploying. Robotaxis have crossed into paid service — half a million paid rides a week at the leader — but the best data also shows a plateau nobody announced: that ride count has been flat for six months while the operator expanded from 10 metros to 14. Tesla reports miles, not rides, and they're flat too.
3. Promised. General-purpose humanoids: the models are improving fast, the hardware cost is falling (a Chinese humanoid lists at $13,500), and the money has arrived — roughly $78 billion of private-company valuations on undisclosed or nil revenue. Against that: about two thousand full-size humanoids sold and paid for by the one company that reports audited numbers; Tesla's own count, "several hundred units… for learning, not productive tasks"; and only ~10% of last year's production doing real work.

Valuation, in one sentence: the companies that ship robots at scale trade at 22–34 times next year's earnings; Tesla trades at 186 times and the newly listed Unitree at 259 — 623.7× on the earnings it has actually reported. Those two prices are the humanoid premium.

The three numbers that matter

What would change our mind

The full report scores eight measurable markers every quarter — deployed growth, verified humanoid deployments, unit cost, autonomy scale, private-capital behavior, platform pull-through, China's cost curve, and thematic fund flows. Today's reading: one green (the deployed business), one red (private capital — step-ups on no deployments, the froth signal firing), six amber. We publish the scorecard so readers can check us, not trust us.

What's in the full report

The value-chain map · the three segments graded contracted / operating / announced · the humanoid roster with every claim marked verified or claimed · Tesla's Optimus timeline in Tesla's own words · what the buyers say (a 23% satisfaction rate) · the forecast-spread table · 12 core-company profiles with what each price assumes · the fund purity table (the largest "robotics" funds share almost nothing but Nvidia; all US humanoid funds together hold ~$490M) · the client Q&A · full sourcing.

Disclosures. Impersonal research, identical for every purchaser and tailored to no one; not individualised advice, and Finnacl Research is not an investment adviser (publisher's exclusion, Advisers Act §202(a)(11)(D)). The tiers differ in seats and redistribution rights, not in content. No statement about what any reader should hold, no recommended allocation, no forecast. The author's holdings are disclosed at the front of the report; the Churchill Capital XI position is frozen while the report is live. No issuer payment, no advertising, no trading against published views, and no trade in any security named here in the thirty days before or after publication. Full disclosures in §13 of the report.