Robot labor estimates range from $2 to more than $35 an hour, first reported by Fortune, because they use different cost and utilization assumptions. That gap separates low running-cost estimates from full deployments and shapes whether physical AI can move beyond costly pilots into a labor market measured in trillions of dollars.
Roland Berger’s $2 figure covers energy, software, network and scheduled maintenance, but excludes buying or leasing the machine. It beats average U.S. warehouse wages only when robots run 60-70% of the time.
The figures measure different things: some describe running costs at volume, while others include equipment and integration or reflect early deployments with low utilization and extensive human supervision. Productive hours connect those estimates to a robot’s effective cost.
A $90,000 industrial robot amortized over five years contributes about $9 per operating hour at 2,000 working hours a year; maintenance and other expenses could bring the total near $12. At 1,000 hours a year, the hardware cost per productive hour doubles.
JPMorgan estimates that it currently takes about two humanoids to match one human worker’s output, doubling the effective hourly cost. Productivity per hour improves with software, while cost per hour improves only with operating hours.
Agility Robotics’ S-4, filed ahead of its merger with Churchill Capital Corp XI, says Digit robots have logged more than 65,000 hours across nine customer sites, including GXO, Schaeffler and Toyota Motor Manufacturing Canada. The filing lists $1.78 million in 2025 net sales against a $140.2 million operating loss, plus $300 million in milestone-contingent orders from one customer.
Agility’s model puts its next robot at roughly $400,000 over five years: about $40 an hour for one 2,000-hour annual shift and close to $20 for two shifts. Against Agility’s $30 fully loaded worker benchmark, the robot becomes cheaper only after more than 2,700 hours a year.
High utilization still depends on reliable work. Factories and warehouses bring changing packaging, unexpected object positions and people entering aisles, which can turn a demonstration task into repeated exceptions.
Labor-as-a-Service shifts the purchase from equipment to outcomes such as pallets moved or packages sorted, with providers taking greater responsibility for deployment, maintenance, software and uptime. A provider paid for completed work depends on keeping its machines productive.