+First time on the top traders desk?
The desk runs its own numbers and publishes its calls next to the account that trades them. The short route:
by Is physical AI investable?
Every few weeks a new humanoid demo goes viral and the question returns: is physical AI actually investable? The answer is yes, but not the way most people try to play it.
- The desk's answer: physical AI is investable, but not the way most people play it. Robotics is a bet on the supply chain, energy, compute, sensors, control systems and installed base, not the humanoid shell everyone films.
- The durable trade is the picks and shovels that take a cut of the whole physical-AI economy: lithography, foundries, GPUs, EDA, power and cooling, and industrial/medical robotics with real installed bases.
- Unit economics decide it. The trigger is never the robot video but whether a company can ship, service and retain users without unit economics exploding.
- The desk map: public core (semis, EDA, power, established robotics), private upside (defense, autonomy, industrial), and the avoid bucket (generic humanoid startups without real deployments).
- The TT desk call: buy the supply chain that earns from the whole economy, not the humanoid startup chasing the next viral clip. Being early and right on physical AI pays the most and is hardest to time.
Not a bet on the robot
Robotics is not a bet on a robot. It is a bet on the supply chain, energy, compute, sensors, control systems, simulation and installed base. The humanoid shell everyone films is the least investable part.
Demos are easy. Ops are the moat.
Own the picks and shovels
The durable trade is the supply chain that takes a cut of the entire physical-AI economy: lithography, foundries, GPUs, EDA, power and cooling, and the industrial/medical robotics with real installed bases.
- Public core: semis, EDA, power, established robotics
- Private upside: defense, autonomy, industrial
- Avoid: generic humanoid startups without real deployments
Unit economics decide it
The trigger is never the robot video; it is whether a company can ship, service and retain users without unit economics exploding. Paid pilots, on-device reliability and real deployment separate the trade from the narrative.
The TT desk thoughts
Being early and right on physical AI pays the most and is hardest to time. Buy the supply chain that earns from the whole economy, not the humanoid startup chasing the next viral clip.
What would change our mind
An assembler capturing supply-chain margin through integration, or a component tier commoditising faster than the robots ship. The position is that the picks-and-shovels layer keeps the economics; either of those moves it.
Keep reading
Physical AI & Robotics Research: the full framework · All alpha