Humanoid robot costs to fall 50% as physical AI scales

While the first wave of artificial intelligence focused on disrupting knowledge workers—copywriters, bookkeepers, and junior developers—the next frontier is the physical world. Physical AI represents a seismic shift where machines no longer just process text but navigate complex environments. This movement targets the "skilled trades" sector, which encompasses warehouse pickers, delivery drivers, and factory assemblers, representing a significantly larger portion of the global workforce than office-based roles.

Vision language action models bridge the reality gap

Historically, industrial robots were rigid, requiring expensive reprogramming for every minor task adjustment. The breakthrough lies in Vision Language Action Model (VLA). These models allow robots to treat physical movements as a form of language, enabling them to generalize and respond to spoken instructions like "pick up the red box" without prior specific training. This capability moves robotics from a fixed program to an adaptable intelligence.

Falling hardware costs and the demographic imperative

Economic feasibility is arriving faster than many anticipated. Bank of America research indicates the bill of materials for humanoid robots will drop from $35,000 in 2025 to under $17,000 by 2030. Concurrently, a declining US labor participation rate—falling from 67% in the late 90s to 62% today—creates a structural labor shortage that only automation can fill at scale. As manufacturing reshores to high-cost regions, companies like Apple and TSMC are making billion-dollar bets that necessitate robotic efficiency.

Strategic positioning in the robotics value chain

Investors should view this landscape through a "picks and shovels" lens. While humanoid startups like Figure AI carry high valuations and execution risks, the underlying infrastructure is more resilient. Nvidia has positioned itself as the foundational platform for almost every robot company. Established industrial giants such as Fanuc, ABB, and Siemens offer lower-risk exposure, as they are already profitable and upgrading their hardware with AI capabilities. Diversification through thematic funds like the WisdomTree Physical AI, Humanoids and Drones ETF provides a safeguard against the uncertainty of which specific hardware manufacturer will ultimately dominate the market.