The most revealing number in the new global robot tally is not the one attached to humanoids. It is 117,500: the number of transport and logistics robots sold in 2025. According to the International Federation of Robotics (IFR), those practical machines accounted for 47% of professional service-robot sales. Full-size humanoids sold about 7,000 units.

That gap matters because the public image of robotics is almost upside down. The largest established markets are for narrower machines. These systems move goods, clean floors, carry instruments or perform one defined medical task. The robot economy is growing by fitting machines to workflows—not by solving the human body all at once.

A Boom with Several Different Denominators

The IFR says shipments of professional service robots rose 24% to almost 250,000 units in 2025. Consumer service robots, a much larger category dominated by domestic machines, rose 37% to nearly 34.2 million units. Medical-robot sales increased 19%, while surgical robots alone reached 8,400 units, up 57%.

Those percentages cannot be added into a single picture of “the robot market.” A robotic vacuum, an autonomous warehouse cart and a surgical platform differ in price, risk, operating environment and what counts as a sale. The IFR figures are supplied by participating manufacturers and summarized in its paid World Robotics 2026 report; the public release gives unit counts, not a complete independent audit of deployments or utilization.

Still, the mix is useful. It shows where buyers already find a business case. Transport robots work in mapped spaces, repeat routes and move standardized loads. Cleaning robots can cover large floors on schedules. A surgical robot operates inside a tightly managed clinical system with trained staff. Each job reduces the number of surprises the machine must handle.

Why Wheels Win before Legs

A wheeled platform avoids one of robotics’ hardest problems: keeping a tall, articulated body balanced while it moves through clutter and safely handles unfamiliar objects. Wheels trade versatility for stability. Warehouses trade open-ended human environments for lanes, markers, access rules and predictable inventory.

That is not a cheat. It is product design. A useful robot is a combination of machine and environment. The more the surroundings can be structured, the less intelligence, sensing and mechanical dexterity the robot needs to deliver reliable work.

The IFR’s business-model note reinforces that point. In its discussion of transport and logistics, it says robot-as-a-service has become the dominant US business model. Instead of buying a machine outright, a customer pays for access or completed work. That can lower the initial barrier, but it also shifts attention from a spectacular demonstration to uptime, maintenance and whether the robot completes enough useful tasks to justify the recurring fee.

The Humanoid Number Needs Context

About 7,000 full-size humanoids above 140 centimeters were sold globally in 2025 for professional applications beyond research and entertainment, according to the IFR. That is no longer zero, but it is not evidence that general-purpose humanoid labor has arrived.

The federation itself stresses that current deployments remain specialized and often depend on human teleoperation. It lists safety standards, training and maintenance costs, manipulation, programming and the absence of a strong business case among the remaining barriers.

Teleoperation is especially important. A robot that appears autonomous may have a remote human handling difficult moments, collecting training data or rescuing failed tasks. That can be a sensible development strategy. It is simply different from a robot that independently handles the full range of situations its body suggests.

Watch the Work, Not the Shape

The next decisive evidence will not be another choreographed humanoid video. It will be operating data: hours worked without intervention, tasks completed, failures, maintenance cost, worker safety and the share of difficult cases handed to people.

The same standard applies to less glamorous machines. A warehouse cart that moves boxes every night can be commercially transformative even if it never folds laundry or climbs stairs. The 2025 market figures suggest that this quieter definition of progress is already winning.

Keep Reading

Production note: Vastkind analyzed the IFR’s public World Robotics 2026 Service Robots release and its disclosed category definitions. We did not independently audit the underlying manufacturer submissions or purchase the full report.