What 86 Chinese executives really think about humanoid robots
Close but no cigar?
Morgan Stanley asked 86 Chinese C-suite executives whether they would buy humanoid robots. The results are the clearest picture yet of a market that wants to say yes but can't quite get there.
The bank's first AlphaWise survey of Chinese industrial buyers, conducted in late 2025 and led by China industrials analyst Sheng Zhong, found that 62% of respondents said they were likely to adopt humanoid robots within 3 years. That is a striking number.
The problem is everything else they said. Basically, the robots aren't ready yet.
The satisfaction gap
Only 23% of respondents said they were satisfied with the humanoid robots currently available.
That means more than 3 in 4 potential buyers have looked at the current crop of products and found them wanting. The specific complaints?
- insufficient dexterity
- limited functionality
- prices that do not reflect the capability
Analyst Sheng Zhong noted that the 62% adoption likelihood "may be optimistic," since the survey sample skewed toward large enterprises that already use robotics. These are the friendliest possible buyers.
If 77% of them are unsatisfied with current products, the picture for the broader market is worse.
The price problem
Chinese industrial buyers have a clear price in mind: RMB 100,000 to 200,000 per unit, roughly $14,000 to $28,000. That is the range where 92% of respondents said mass adoption becomes viable.
Below $28,000, companies will seriously consider buying. Above it, most won't.
Where are humanoid robots actually priced?
Entry-level research-grade robots from Unitree start around $4,900 to $13,500, but these are not industrial-grade machines with the dexterity and reliability manufacturing buyers need. The industrial-capable units that could do real factory work, things like UBTECH's Walker S2, run $68,000 to $120,000. Fourier's GR series goes $100,000 to $170,000.
Ergo: the gap between what buyers want to pay and what capable robots cost is still significant.
Morgan Stanley estimates the average price could fall to roughly $21,000 by the end of the decade, down from about $46,000 now. In business terms, that's an eternity away. So the big question is whether that price curve bends fast enough to catch the window of opportunity before industrial buyers lose patience and automate through other means.
Only 10% are actually doing anything
Despite the stated interest, only about 10% of respondents are currently evaluating or launching pilot projects.
The other 90% are watching. This is the gap between aspiration and action that defines the early market phase. Companies believe the technology is coming, plan to adopt it eventually, but are waiting for the products to get good enough to justify moving.
In other words: the ball is in the robot makers' courts to improve quickly enough in both capability and pricing.
Morgan Stanley also found that current humanoid robots operate at roughly 20% to 30% of human worker efficiency ... the root cause of the ROI problem. A robot at 25% human efficiency, priced at $68,000, on a 3-year payback requirement, simply does not pencil out against a Chinese manufacturing worker at typical wage rates.
The math only starts working when efficiency rises and price falls simultaneously.
What buyers actually want: dexterity & self-learning
The survey asked about capability priorities.
The top demands were dexterity (cited by 57% of respondents) and self-learning capability (49%). Both of these are among the hardest problems in the field. Dexterity requires fine motor control that current actuators struggle to deliver consistently in unstructured environments. Self-learning means generalization, the ability to handle tasks the robot was not explicitly trained on, and that is the core unsolved problem of physical AI.
(Note: Standard Bots just unveiled something very interesting that solves this problem.)
Unitree was identified as the most engaged brand among respondents, followed by DeepRobotics, UBTECH, and Midea.
The Unitree lead makes sense: the company has the most visible products at the lowest accessible price points and the highest volume of real-world deployment data. Whether its industrial-capable units can clear the dexterity bar that buyers are demanding is a different question.
The longer-term view
Respondents believe 11% of jobs could be replaced by robots within 5 years and 28% within 10 years. Morgan Stanley's broader forecast projects the Chinese humanoid market at $2 billion in 2026, growing to $15 billion and 446,000 annual units by 2030. The firm has raised its 2026 China shipment forecast three times this year: from 14,000 to 28,000 to 50,000 units.
Those forecast upgrades reflect real commercial traction. But they also reflect a market still heavily weighted toward pilots, demos, government research programs, and speculative purchases rather than sustained industrial deployment. The survey data is the gut-punch counterweight to the shipment forecasts: intent is high, products are not ready, and the price-capability intersection that triggers mass adoption has not arrived yet.
The companies that close that gap first, on price and on real dexterity at scale, will capture a market that is clearly waiting to be captured.
The 62% are ready to buy. They are just waiting for something worth buying.
Sources
Written by
I tell stories with data and build startups. I've raised angel and VC funds, led software companies and startups, and chronicled the rise of the mobile economy. Currently I write for Forbes, host several podcasts, and am starting up HumanoidMegaHub. I was TUNE's mobile economist and Singular's VP of Insights, built the research team at VentureBeat, and managed teams creating software for partners like Intel and Disney. In past lives I built websites, e-learning platforms, and mobile apps, and, a long time ago, was named to Folio's top 100 of the media industry's "most innovative entrepreneurs and market shaker-uppers.”