Physical AI' is here. UBS shares 3 picks-and-shovels sectors of the market to play the trade.

UBS says “Physical AI” is moving from theory to deployable hardware, driven by humanoid robots that can use in-context learning—learning how to perform tasks by observing actions. Ulrike Hoffmann-Burchardi, UBS’s global head of equities, argued that the industry is nearing a shift away from preprogrammed automation toward “genuine physical autonomy.” In a note to clients, she highlighted three “picks-and-shovels” areas linked to AI robotics: actuators and precision gears for fine finger movement and balance; sensors and perception layers, including cameras and “skin” that can measure pressure and friction; and edge computing and power infrastructure to process data and run robots longer. UBS did not name specific companies in every segment, but cited examples including Rockwell Automation, ABB, Synaptics, Ouster, Luminar, Nvidia, Arm and AMD.