Benchmarks

Benchmarks

Georgian + NewtonX Physical AI Survey, Wave 1 · n=100 respondents (85 buyers / 15 builders) · Fielded July 24 to Aug. 21, 2026 · Reporting base: Buyers, n=85 unless a benchmark states otherwise

  1. H1 — Less Fragile Each Time, Not a Better DemoProgress in physical AI shows up as each deployment being less fragile than the last, not as a better demo.
  2. H2 — Interventions Can Be ManufacturedIntervention rate is useful but incomplete.
  3. H3 — A Systems Market, Not a Model MarketPhysical AI is not simply a model market.
  4. H4 — Evaluation Discipline as the SignalIn physical AI, the ability to measure task success, degradation and real-world robustness is often as important as the model itself.
  5. H5 — Judge It as a Learning EngineEarly deployments are better understood as learning engines than finished products.