IDC estimates that about 18,000 humanoid robots shipped worldwide in 2025, a 508% year-over-year increase, with roughly $440 million in reported sales. China appears to lead in shipment volume and the breadth of its robotics industry. But shipments are not the same as productive robots at work: most reported deployments were demonstrations, education, data collection, guided tours, or validation—not proven, large-scale autonomous labor.
What does the 508% figure actually measure?
It is growth in estimated unit shipments, not a 508% increase in revenue, market value, installed robots, profits, or working hours. IDC’s 2025 estimate is approximately 18,000 global shipments and around $440 million in sales. The figures are reported in IDC’s 2026 commercialization analysis and summarized by China’s statistical portal in its account of the IDC report.
If 18,000 units represent a conventional 508% increase over the prior year, the implied 2024 base is about 2,960 units: the 2025 total is roughly 6.08 times the previous year’s, because a 508% increase means adding 5.08 times the baseline. That back-calculation is approximate, not a separately reported count. It shows why the percentage is striking while the absolute market remains small.
Dividing the reported $440 million in sales by 18,000 units gives an arithmetic average of about $24,000 per shipment. That is not a standard selling price: the calculation combines different robot types and sales and says nothing about a specific model’s price, integration costs, or vendor profit. The sales figure is summarized by China Aid Daily’s report on IDC’s figures.
#1 Best Overall
Why shipment estimates do not settle how many robots are working
There is no single established count in the available figures for humanoids that are operating autonomously in sustained, productive service. Shipments may include hardware bought for research, developer experimentation, demonstrations, or short pilots; a shipment count does not establish that a robot is still in service, completes tasks without frequent human help, or earns a commercial return. Nor do the estimates provide a consistent breakdown of full-size bipeds versus smaller or otherwise differently classified platforms.
Estimates also differ by firm and possibly by definition or coverage. IDC puts 2025 global shipments at about 18,000. An AP report citing Omdia-based estimates gives a lower global range of about 13,000 to 16,000 and says AgiBot and Unitree each shipped more than 5,000 units. Those are separate estimates, not figures that should be averaged or silently combined. The public accounts do not resolve exactly how much of the gap reflects definitions, reporting periods, or inclusion of adjacent robot categories.
| Measure | Reported estimate | What it does—and does not—show |
|---|---|---|
| Global shipments, IDC | About 18,000 units in 2025; up 508% year over year | Estimated units shipped, not a verified count of productive or autonomous workers. |
| Global shipments, Omdia-based AP reporting | About 13,000–16,000 units in 2025 | A separate estimate; its difference from IDC is not fully reconciled in the available accounts. |
| Reported sales, IDC figure | About $440 million in 2025 | Sales revenue, not profit, company valuation, or total economic impact. |
The shipment estimates and the deployment breakdown are reported by IDC; the alternative range appears in AP’s report citing Omdia.
China leads in volume, but the lead has boundaries
IDC’s account ranks Chinese company AgiBot first in 2025 shipments, at approximately 5,200 units, including about 1,300 full-size humanoids. The distinction matters: the total is not a count of 5,200 full-size bipedal robots. CGTN’s summary of IDC’s ranking gives the full-size breakdown. AgiBot also announced the results itself; its company announcement is useful for its own claims, but self-reporting is not independent confirmation.
Recommended Free Tools
Rank #2
- Ideal for Gifting
- Ideal for a bookworm
- Compact for travelling
China’s Ministry of Industry and Information Technology said the country released more than 330 humanoid-robot products in 2025 and had more than 140 manufacturers, according to AP’s reporting citing official data. These figures describe breadth, not how many firms have meaningful recurring sales, reliable production, or commercially successful robots. A large vendor field can speed experimentation and competition while also foreshadowing consolidation.
Companies to watch in China
- AgiBot: IDC’s reported shipment leader. Its position is significant for volume, but shipment rankings alone do not establish long-term autonomy, customer returns, or technical superiority across every task.
- Unitree Robotics: Known for comparatively accessible robotics platforms and high visibility. Its official site is a starting point for product information, but model availability, configuration, and total procurement costs depend on buyer and region.
- UBTECH: An enterprise- and factory-oriented example. Its official site describes the company’s offerings; a factory demonstration or pilot is not evidence by itself of repeatable production economics.
- Other names: Leju Robotics, Fourier Intelligence, EngineAI, LimX Dynamics, Galbot, Xpeng Robotics, and Deep Robotics operate across different parts of the broader robotics landscape. They should not be treated as interchangeable manufacturers of the same kind of full-size humanoid; some focus on research platforms, embodied-AI systems, or adjacent robot forms.
Competitors outside China
Companies and organizations including Tesla Optimus, Figure AI, Apptronik, Agility Robotics, Boston Dynamics, Sanctuary AI, 1X, and PAL Robotics are part of the international field. Funding, a prototype, a public demonstration, a factory pilot, a shipment, and a paying customer are different kinds of evidence. None alone establishes the most commercially capable vendor, and a shipment ranking does not decide who leads in a particular area such as locomotion, manipulation, safety, or factory integration.
Where humanoids are being used—and what remains experimental
IDC reports that more than 85% of 2025 deployments were concentrated in performances, education, data collection, guided-tour services, and related demonstration or validation settings. The exact denominator and category boundaries matter, but the breakdown is a strong warning against reading shipment growth as evidence of factory-wide labor replacement. It points to a market that is selling hardware and building experience, with widespread productive deployment still unproven.
More accessible uses today
- Education, university research, and developer experimentation.
- Demonstrations, exhibitions, entertainment, and guided tours.
- Collecting demonstrations and other data for robot learning and manipulation research.
- Controlled inspection or validation tasks, where environments and expectations can be tightly managed.
Industrial tasks under development
Companies are exploring moving bins or parts, pick-and-place, machine tending, basic assembly assistance, warehouse handling, repetitive loading and unloading, and factory inspection. A successful demonstration can show that a robot performed a task under particular conditions; it does not establish that the same system can meet a production line’s speed, safety, uptime, and cost requirements over time.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Claims that remain speculative
Broad household labor, reliable elder care, open-ended construction, fully autonomous warehouse operations, and replacing workers across varied jobs require much more than a robot completing a staged task once. A commercially meaningful test asks whether it can work repeatedly, at acceptable speed, without constant teleoperation, recover safely from mistakes, and remain economical as lighting, objects, layouts, and other conditions change.
Why China has an early volume advantage
No single factor explains China’s position. The advantage is better understood as the combination of a manufacturing base, a large domestic market, public and private investment, and a crowded field of companies able to test hardware quickly.
- Manufacturing and electronics supply chains: Dense access to factories and component suppliers can shorten prototyping and iteration cycles. China also has an established industrial-robot and automation ecosystem to draw on.
- Domestic deployment opportunities: Large manufacturing and logistics sectors offer potential sites for trials. Companies and institutions willing to buy early hardware can support research, data collection, and product refinement even before a robot is ready for unsupervised production.
- Policy support and public demonstrations: Government funding, industrial policy, subsidies, and showcase programs can encourage investment and public visibility. Product launches and manufacturer counts, however, are not proof of productive output.
- Hardware cost competition: Vendors such as Unitree have helped make some platforms more accessible to developers and researchers. Lower hardware cost can broaden experimentation, but it does not automatically reduce the full cost of operating a robot safely at a worksite.
- Competition at scale: Many vendors create more experiments and approaches. They also make it harder to distinguish durable businesses from companies whose products may not find repeat buyers.
These factors support the conclusion that China has an early lead in reported volume and industrial ecosystem breadth. They do not prove it leads in every dimension of performance, autonomy, or profitability.
Why build a robot in human form?
The strongest case is compatibility with places people already use. Stairs, doors, shelves, carts, tools, and workstations are typically designed around human bodies. A machine that can move through those spaces and handle existing equipment might be installed without rebuilding a facility around fixed automation.
Rank #4
The counterargument is practical: wheels are generally more efficient on flat floors, and a robot arm, conveyor, or specialized machine can be faster and more reliable for a narrowly defined task. Bipedal balance adds mechanical and software complexity; human-like hands are difficult to control consistently. Humanoid shape is a design option, not evidence that a humanoid is the best machine for a particular job.
The technology stack is more than an AI model
A useful way to understand a humanoid is as a body, a control system, and a software stack that must work together. The hardware includes actuators and gearboxes, motors and controllers, batteries and thermal management, and cameras, depth sensors, force sensors, and inertial measurement units. The software must turn sensor readings into a reliable sense of position and surroundings, then plan and control movement while balancing and manipulating objects.
- Movement and manipulation: Whole-body control, balance, motion planning, and grasping must work together. Failure to recover from a blocked path or dropped object can turn a simple task into a human intervention.
- Learning and perception: Vision-language-action models, task-specific data, simulation, and sim-to-real transfer may help robots interpret instructions and adapt. A language model alone does not provide dependable physical skill.
- Human input and fleet operations: Teleoperation and demonstrations can teach or assist a system, while fleet-management tools and software updates help coordinate robots. A task that appears autonomous may still rely on remote or on-site human support.
- Safety and reliability: Emergency stops, safe behavior around people, maintenance, spare parts, and secure network access are part of deployment—not optional add-ons.
The economics: hardware is only one line of the bill
A buyer evaluating a robot needs task-level total cost of ownership, not just a list price. Relevant costs include integration, maintenance, battery replacement, safety equipment, insurance, human supervision, downtime, and any data or network infrastructure. The relevant output is cost per reliably completed task or unit of production, compared with people and with other automation—not the purchase price in isolation.
A low-cost research platform can be valuable for education, motion experiments, algorithm development, and vision research while still being unsuitable for heavy loads, hazardous work, high-uptime production, or unsupervised industrial operation. Conversely, a more capable enterprise system may require substantial integration and support. Public demonstrations rarely supply enough information to calculate the economics for a specific factory or workplace.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteBest Value
- It can be a gift option
- Comes with secure packaging
- Helpful in various ways
For a company considering deployment, the comparison should include conventional arms, autonomous mobile robots, conveyors, machine vision, and simpler process changes. Humanoids are most compelling where tasks vary, both mobility and manipulation are needed, and the cost of modifying existing facilities is high. Even there, the alternative may be cheaper or more dependable.
What evidence would show that the market has moved beyond demonstrations?
Shipment volume is a useful early signal of production and buyer interest, but commercial maturity requires evidence that customers can get repeatable results. A more informative scorecard for China—and for competitors elsewhere—would track:
- Repeat orders from independent enterprise customers, rather than one-off purchases or showcase programs.
- Units operating outside demonstrations and the number of productive hours they complete.
- Task completion rates, operating speed, and the frequency of human intervention or teleoperation.
- Full-shift endurance, maintenance intervals, failure recovery, and downtime in real working conditions.
- Production and deployment costs alongside output per hour and cost per completed task.
- Safety certifications, liability arrangements, insurance, spare-parts availability, repair networks, and support across export markets.
IDC forecasts that manufacturing deployment growth will exceed 200% in 2026. That is a forecast, not an observed result; it should be judged against subsequent deployment evidence, rather than treated as proof that scale has already arrived.
What the 2025 surge means for workers and investors
The growth shows that humanoid robotics has moved beyond a handful of lab prototypes into a bigger market for hardware, research, and trials. It does not establish that humanoids are already cheaper than human workers, that vendors are profitable, or that broad job replacement is imminent. In the near term, robots may assist with constrained, repetitive tasks; the pace and extent of any labor impact depend on reliable autonomy, task economics, safety, and actual deployment—not on shipment growth alone.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →For investors and financially curious readers, a headline shipment rate is not a substitute for evidence about a company’s revenue quality, margins, repeat customers, service obligations, and ability to turn pilots into sustained contracts. The same distinction matters for manufacturers deciding whether to purchase a platform: an exciting demonstration can justify a trial, but not a claim that a robot is ready to replace a production process.
IDC’s reported 2025 shipments make the 508% headline real as a unit-growth estimate. China’s early advantage is clearest in shipment volume, vendor breadth, and industrial ecosystem scale; whether that becomes durable commercial leadership depends on robots proving they can work safely, autonomously, and economically at scale.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
