Theses›Robotics & physical AI›Hazardous-environment robotics

Hazardous-environment robotics

Hazardous-environment field robotics: machines bought to enter places a person cannot survive or cannot fit

  • Bear
  • Early
  • $5.5B by 2035

The callA real, growing market that doesn't make venture-scale winners: value scatters by hazard class and the inspection houses keep the customer.

D1D2R1R2R3C1C2foragainst D1D2R1R2R3C1C2foragainst
Fig. 1The balance. Drivers and catalysts against restraints, each block sized by its weight; catalysts hang dashed because they have not happened yet, sized by their odds. It tips toward the restraints.
  1. D1Ageing refineries, mines and nuclear sites run short of people willing to enter a 200C vessel or a reactor cell; nuclear decommissioning alone is about $10B a year by 2030.
  2. D2Standing beyond-line-of-sight rules (FAA Part 108, Europe's U-space) would turn a $30k drone flown in by a crew into a $180k resident installation the asset owner buys.
  3. R1Every hazard needs a different machine, so no common architecture lets one winner take the category.
  4. R2The inspection incumbents own the customer: Previan sold Eddyfi to ESAB for $1.45B, buying the house, not a robot maker.
  5. R3Exits price at trade-sale multiples (Boston Dynamics' 80% sale to Hyundai valued it at $1.1B), and the leading company would exit at about €867M, short of €1B.
  6. C1FAA Part 108 final rule published, turning inspection robots into resident assets flying scheduled missions; it reached White House review in July 2026.end-2027
  7. C2EASA adopting or withdrawing the U-space light amendment, the European half of the standing-rules shift.30 Jun 2027
Rests onthe beliefs underneath
Supported by · confidence
What would prove itdated tests, open and resolved
20272028202920302031today123456predictionswatch 20272028202920302031today123456predictionswatch
Fig. 2Each ring is a dated test: dashed while open, struck through when overdue, ticked or crossed once resolved.
  1. 1EASA adopting or withdrawing U-space light.Jun 2027 · open
  2. 2The FAA Part 108 BVLOS final rule is published in the Federal Register by end-2027.Dec 2027 · open
  3. 3Part 108 in the Federal Register.Dec 2027 · open
  4. 4By end-2029 an incumbent NDT or inspection group acquires a European hazardous-environment robot vendor at under 4x trailing revenue.Dec 2029 · open
  5. 5By end-2030 at least one civil hazardous-environment field-robotics vendor, excluding defence suppliers and excluding human-inspector services firms, reports more than $250M of annual revenue.Dec 2030 · open
  6. 6One civil hazard-robot vendor above $250M of revenue (Gecko nearest at about $100M); an inspection group buying a European maker under 4x revenue by 2029.Dec 2030 · open
Expressionslisted, private, technologies
  • BKRBaker Hughes Company
    • fair
  • MGMistras Group, Inc.
    • cheap
  • TDYTeledyne Technologies
    • fair
     Industrial imaging incl.
  • 005380.KSBoston Dynamics
    • not screened
     Sells the Spot legged robot for industrial inspection, nuclear decommissioning and CBRNE work to plant operators (e.g. Sellafield); ~80% owned by Hyundai Motor Group (parent ticker).
  • ESABEddyfi Technologies
    • not screened
     Makes VersaTrax inspection crawlers and remote visual inspection robots for pipes, tanks and nuclear sites; acquired by ESAB Corporation in 2026 (parent ticker).
Private · 6 private companies, 4 of them in Switzerland.

RAROG · ANYbotics · Voliro · Square Robot · Korial (formerly Energy Robotics) · RoBoa

Technologies
Who has already bet on itprivate rounds, from public sources

$200M raised by 5 companies, $115M of it in the last three years. Wave building: Money is arriving: recent rounds and tier-1 funds are in, but no runaway leader yet.

Backed by a top-tier venture firm: ANYbotics, Voliro, Korial (formerly Energy Robotics). Backed by a strategic investor: ANYbotics.

What changedthe record, newest first
  • 16 Aug 2026Hazardous-environment field robotics is roughly four times bigger than a first pass suggested and still does not clear a venture bar, because value fragments across hazard classes and the incumbent inspection houses are buying the robot layer rather than being displaced by it.