Physical AI — AI that senses, reasons, and acts in the physical world — is projected to grow from roughly $5 billion in 2025 to $82.8 billion by 2035.

At a recent industry panel, SVRI joined other manufacturing and technology leaders to unpack what that growth means for the semiconductor ecosystem that has to build it — from advanced packaging to the compute and sensing stack every physical-AI deployment depends on.

The Second AI Wave

The first AI wave was defined by language and reasoning — models that could read, write, and analyze. The second wave, Physical AI, is defined by action: systems that perceive the physical world and operate within it, from robotics on the fab floor to autonomous inspection and material handling. SVRI views this shift not as an incremental upgrade but as a new category of demand for the semiconductor industry to serve.

Physical AI and the Semiconductor Ecosystem

Physical AI sits at an unusual intersection of semiconductor demand, industrial robotics, and enterprise software. Every Physical AI deployment — a warehouse robot, an autonomous inspection arm, an AI-driven manufacturing cell — depends on a stack of compute, sensing, and advanced packaging that has to be manufactured, and manufactured reliably at scale. That puts semiconductor manufacturers in a position they haven't fully occupied before: not just a supplier to AI, but critical infrastructure for a physical economy being rebuilt around it.

Physical AI Market Size: 2025 vs. 2035 ($B)

Industry Discussion at GSISS

Discussion at the GSISS advisory panel centered on what this means in practice for manufacturers already stretched by AI-driven demand: where to prioritize capacity, how to plan for a category of customer — robotics and physical-AI companies — that didn't meaningfully exist in prior investment cycles, and how packaging and compute intelligence need to evolve together rather than as separate roadmaps.

Looking Ahead

With Physical AI markets compounding at over 30% annually across every major region tracked, semiconductor manufacturers that build the planning and benchmarking intelligence to serve this demand early — rather than reactively — will be best positioned as the second AI wave moves from pilot deployments to industrial scale.

Key Metrics & Projections

  • Global market: $5B+ (2025) → $82.8B (2035), a 32.8% CAGR.
  • U.S. market: $2.3B (2025) → $27.3B (2035), a 31.7% CAGR.
  • Fastest-growing region: Japan, at $307.3M (2025) → $6.8B (2035), 36.2% CAGR.
  • Public investment: $52.7B in CHIPS Act funding underpins the compute capacity this depends on.