Physical AI, All Layers (practice)
Parth@parth
−9.5%
since Oct 7
The idea
I believe AI is moving off screens and into the physical world, with robots, autonomous vehicles and automated factories shifting from pilots to real deployment over the next five years. To express it I built four layers: the chips and sensors giving machines sight and reasoning, the companies building the robots and vehicles themselves, the sensing and motion suppliers underneath, and the first operators actually deploying this at scale. Big, established names anchor the core exposure while smaller, higher-upside positions stay sized small, and I keep 10% in cash to add on pullbacks.
What it owns
- The chips that let machines think, a layer every robot needs no matter which manufacturer wins. · NVIDIAIt supplies the AI computing platform most robots will run on, so widespread deployment of machines in the physical world flows through its data centers and chip sales., Taiwan SemiconductorIt manufactures the advanced chips for nearly every other company in this basket, so it wins across all branches of physical AI regardless of which maker leads., Arm HoldingsIt designs the chip architecture that low-power devices run on, so more deployed robots and autonomous machines means more licensing revenue almost automatically., QualcommIt specializes in low-power AI chips for vehicles and machines that process information without constant connection to remote data centers., AmbarellaIt makes chips that let machines process video and understand images locally on the device, essential for robots and autonomous vehicles to see without cloud dependence.26%
- The companies actually building robots, from humanoids to surgical and delivery machines. · TeslaIt is building both humanoid robots and autonomous vehicles, making it the largest single bet that physical AI reaches real-world scale and economic value., SymboticIt builds complete automated systems for warehouse distribution centers, where physical AI is deploying at the largest scale first., Intuitive SurgicalIt pioneered robotic surgery and holds a dominant position in the operating room, proving machines can perform precise physical work that hospitals will pay for at scale., PROCEPT BioRoboticsIt is a smaller surgical robot maker pursuing procedures outside Intuitive Surgical's current focus, betting that robotic surgery expands into new specialties., Serve RoboticsIt operates delivery robots on real sidewalks, a visible test of whether physical AI works reliably outside controlled environments like factories and warehouses.25%
- Sensors, cameras and precision parts that let robots see and move, a bottleneck in scaling physical AI. · TeradyneIt owns Universal Robots, the leading maker of collaborative robot arms that work alongside people on factory floors and automate assembly work., CognexIt leads in machine vision systems that let factory robots see and identify what they are working on, essential infrastructure if automation scales beyond pilots., NovantaIt makes the precision motion and sensing components inside surgical robots and factory automation, a picks-and-shovels position that benefits from all robot deployment., MobileyeIt builds the perception systems that enable autonomous vehicles to see and understand their environment, a core layer of physical AI on roads., OusterIt makes lidar sensors that help robots and autonomous vehicles map and understand the environment around them, sized small given the early stage of lidar adoption., Regal RexnordIt manufactures the motors and actuators that power robot joints and automated equipment, a component every physical machine in this basket requires.21%
- The warehouses, factories and farms putting robots to work first and pocketing the savings. · Rockwell AutomationIt sells the control systems and software that run automated factories, so widespread manufacturing automation flows directly through its order book and revenue., AmazonIt operates more deployed robots than any other company and has acquired multiple robotics makers, proving that large operators view automation as essential to cost savings., Deere & CompanyIt is deploying autonomous tractors and AI-powered crop management tools on farms, extending physical AI into agriculture where human labor is scarce., GXO LogisticsIt operates warehouses for other companies and is piloting humanoid robots in them, providing an early test of whether experimental robots move into sustained commercial use.18%
- Cash10%