Global humanoid robot shipments approached 25,000 units in the first half of 2026, according to IDC data reported by Robotics & Automation News. That is up 432.1 percent year on year. Agibot ranked first globally by shipment volume, ahead of every other maker tracked by IDC.
For anyone evaluating a humanoid platform to buy or pilot, that number matters more than any stage demo. Shipment data measures what a company can actually manufacture, deliver and support at scale. Spec sheets and keynote choreography do not.
Why shipment volume is the signal that counts
Humanoid robotics has spent recent years full of research previews: capable-looking machines that perform scripted tasks in front of cameras, with no public unit count behind them. IDC's H1 2026 figures cut through that. A maker shipping at volume has solved problems that never show up in a demo reel: manufacturing yield, field service, firmware updates across a fleet, and supply chain for actuators and batteries.
Agibot's top ranking puts it in a different category from competitors still running small pilot deployments. That does not mean Agibot's hardware is the most capable on paper. It means Agibot is the maker most likely to have a working support and replacement pipeline if a unit fails in the field, a question that matters more to an operator than any torque spec. Buyers comparing platforms can check current specs and pricing on the humanoid comparison page and track list prices on the price tracker.
The 432 percent figure deserves context, not hype
A 432.1 percent year-on-year increase sounds dramatic, and it is a real inflection point. It is also a growth rate measured from a small base: the humanoid market a year earlier was still mostly pilots and demonstrators. Triple-digit growth off a low starting point is easier to post than it is to sustain once volumes climb much higher. The number confirms that 2026 is the year humanoid shipments left the pilot stage in aggregate. It does not yet tell buyers which individual makers will still be shipping at this pace as the market matures.
Pilots are still real, just a different bet
Not every credible humanoid player is chasing shipment volume. Neura Robotics is expanding its physical AI and cognitive robots into healthcare and life sciences, per Robotics & Automation News, a move into a regulated, high-stakes vertical rather than a push for unit count. That is a legitimate strategy, but it is a different one. A healthcare pilot program is a different kind of proof than a shipment ranking: it says nothing about delivery timelines or fleet support at scale. Buyers should treat the two signals separately: shipment leaders for organizations that need hardware now, specialist pilots for those willing to wait on a narrower, deeper deployment. The best humanoid robots list tracks both categories.
Infrastructure is catching up to the hardware race
Away from unit counts, the industry is also building the plumbing that embodied AI will need at scale. At the Humanoid Robots Summit 2026 in Stuttgart, FMC³ Robotics, Vector Informatik and PEM Motion signed a memorandum of understanding to jointly develop standardized data infrastructure for embodied AI, according to Robotics & Automation News. Standardized data pipelines are unglamorous compared to a shipment ranking, but they are a precondition for any maker trying to train and update fleets of robots across different hardware generations.
There is similar groundwork happening on the software side. NVIDIA's IsaacTeleop converts XR hand tracking and motion controller input into robot commands through a graph-based retargeting engine, per MarkTechPost, a tool aimed at making teleoperation data collection more practical for training robot policies. Separately, Runway has introduced Praxis-1, a world action model for robotics that it plans to release with open weights, per The Robot Report. Open-weight models lower the barrier for smaller makers to build capable robots without each developing a foundation model from scratch.
The Robot Report also flags that the physical AI race will increasingly be shaped by patent strategy, with robotics startups advised to be selective about which IP they protect rather than filing broadly. None of this changes a purchase decision today, but it is part of the groundwork that will decide which makers are still shipping, not just demoing, down the line.
What this means for a purchase decision today
IDC's H1 2026 data gives buyers a genuinely new filter: who is shipping, not who is demoing. Agibot's position at the top of that list is a stronger data point than any single capability claim, because it reflects manufacturing and logistics that are far harder to fake than a video. That said, shipment volume is one input, not the whole decision. A healthcare operator evaluating Neura Robotics' push into life sciences is solving a different problem than a warehouse operator looking for a platform it can deploy in quantity this year.
The practical approach is to separate the two questions. First, which makers have demonstrated they can deliver and support hardware at volume, a list IDC's data now meaningfully narrows. Second, which of those makers' specs and price points fit the actual task. Cross-reference both before committing budget to a platform.