Which humanoid robotics startups are actually making the most money today?
In our humanoid robotics market deck, you will find everything you need to understand the market
SUMMARY
AgiBot is the best answer to which humanoid robotics startups are actually making the most money today, while Unitree has the strongest proven combination of large humanoid revenue and real profitability.
The commercial leaderboard is much more Chinese than the valuation leaderboard. AgiBot, Unitree and UBTECH were already generating roughly $115 million to $150 million of annual humanoid-related revenue each in 2025, while most heavily funded Western names were still far earlier in recognized sales.
AgiBot looks like the current private-company revenue leader because it reported RMB1.05 billion of 2025 company revenue and says it exceeded RMB1 billion again in the first quarter of 2026. That quarterly pace is the clearest sign that the ranking may be moving faster than full-year accounts can capture.
Unitree is the cleaner financial benchmark. It disclosed roughly RMB868 million of 2025 humanoid revenue, RMB1.70 billion of total sales and about RMB288 million of attributable net profit, giving it something most humanoid companies still do not have: scale and profit at the same time.
UBTECH is almost as large in humanoid revenue but follows a very different model. Its robots generate much more revenue per delivered unit than Unitree's, which fits a more industrial, enterprise-heavy and customized deployment strategy, but the company is still losing substantial money.
Price and volume are already separating business models. Unitree has pushed humanoid prices low enough to sell thousands of units, while UBTECH earns far more per robot; Leju has also cut average Kuavo pricing while driving a huge increase in volume.
Factory automation is not yet where most humanoid revenue comes from. Data collection, research, education, entertainment and performance still account for a large share of real sales, which is a useful correction to the endless footage of robots walking around car plants.
Orders can make a company look much bigger than its income statement. Agility Robotics has announced more than $300 million of Digit orders, yet its 2025 net sales were only about $1.8 million, so most of that commercial value still sits in future deployments and multi-year contracts.
The valuation gap is even more extreme. Figure has raised capital at a valuation in the tens of billions while disclosing nothing close to the nine-figure humanoid revenue already reported by the Chinese leaders; investors are pricing the possibility of future scale, not the business that exists today.
The practical ranking is therefore pretty clear. AgiBot appears to lead private-startup revenue, Unitree is the strongest financially proven operator, UBTECH is a fast-growing but loss-making heavyweight, Leju is the clearest second-tier commercial player, and the major U.S. startups still have much more funding and pipeline than recognized robot revenue.
This market map, featured in our humanoid robotics market deck, highlights top companies and startups in the humanoid robotics market
Why does humanoid robot revenue suddenly matter?
Humanoid robotics is finally generating enough real revenue for us to separate companies selling robots from companies mainly selling a future vision.
Counterpoint Research estimated that global humanoid robot sales passed $500 million in 2025 for the first time. That is still a small industry beside conventional industrial robotics, but the change from a year earlier was huge. AgiBot went from roughly RMB60 million of company revenue in 2024 to RMB1.05 billion in 2025. Unitree climbed from RMB393 million to about RMB1.70 billion. UBTECH's full-size humanoid business jumped from RMB35.6 million to RMB820.6 million.
The latest numbers show this was not a one-year burst. AgiBot says revenue exceeded RMB1 billion in the first quarter of 2026 alone. Unitree reported RMB423 million over the same quarter. UBTECH later disclosed RMB590.3 million of full-size humanoid revenue for the first half of 2026, up more than fifteenfold year over year.
We can finally answer a question that was almost impossible to answer two years ago: who is actually selling humanoid robots at meaningful scale?
What should we count as humanoid robotics revenue?
For humanoid robotics, we should rank companies primarily by revenue already recognized from robots and related services, while keeping funding, orders and valuations separate.
That distinction changes the leaderboard dramatically. A company raising $1 billion has convinced investors to finance it. A company announcing $300 million of orders may have customers willing to spend that amount over several years. Neither figure tells us how much money the business is earning today.
Company-wide revenue can also mislead. Unitree still sells quadruped robots. UBTECH sells other types of robots and services. 1X built a commercial business around the wheeled EVE robot before its current push into the bipedal NEO. NEURA Robotics sells several kinds of cognitive and collaborative robots.
Where companies disclose humanoid-specific revenue, we use it. Where they do not, we use total robotics revenue carefully and say what is included. Orders and shipment numbers help us judge where the business is going, but they do not replace actual sales.
As this chart shows, and as featured in our humanoid robotics market deck, search interest in where to buy robots has been rising steadily
Who made the most humanoid robot revenue in 2025?
AgiBot, Unitree and UBTECH formed the clear top tier of humanoid robot revenue in 2025, with everyone else well behind them.
Counterpoint Research estimated AgiBot's humanoid-related sales at more than $140 million and ranked it first globally. AgiBot itself reported RMB1.05 billion of company revenue. Unitree disclosed approximately RMB868 million of humanoid revenue within RMB1.70 billion of total sales. UBTECH generated RMB820.6 million from full-size embodied humanoid products and services.
The exact order between AgiBot and Unitree depends on which definition we use because AgiBot does not publish the same product-level financial breakdown as Unitree. The broader conclusion is much firmer: three Chinese companies had already reached roughly $115 million to $150 million of annual humanoid revenue.
Leju Robotics was much smaller at RMB177.8 million from its Kuavo full-size humanoid line. Agility Robotics, despite being one of the best-known American names, reported only around $1.55 million of robot sales.
| Company | 2025 relevant revenue | What the figure covers | Position |
|---|---|---|---|
| AgiBot | >$140m estimated | Humanoid sales estimate from Counterpoint | Top tier |
| Unitree | RMB868m | Humanoid robots | Top tier |
| UBTECH | RMB820.6m | Full-size humanoid products and services | Top tier |
| Leju Robotics | RMB177.8m | Kuavo full-size humanoids | Second tier |
| Agility Robotics | ~$1.55m | Robot sales | Early commercialization |
Is AgiBot now the biggest private humanoid robotics startup by revenue?
AgiBot currently looks like the largest private humanoid robotics startup by revenue, and its latest sales pace appears to have moved ahead of Unitree's.
AgiBot reported RMB1.05 billion of revenue in 2025, versus roughly RMB60 million in 2024. That is about seventeen times more revenue in one year. The company also shipped more than 5,000 robots during 2025.
Then the pace accelerated again. AgiBot says it generated more than RMB1 billion during the first quarter of 2026, almost matching its entire 2025 revenue in three months. Unitree reported RMB423 million in the same quarter. On that basis, AgiBot was running at more than twice Unitree's quarterly revenue.
Production supports the idea that the gap is commercially meaningful. AgiBot reached its 10,000th general-purpose embodied robot in March and its 15,000th by late June. Omdia had already estimated more than 5,100 AgiBot shipments during 2025.
There is one important limit to how confidently we can rank the two. Unitree's numbers come through detailed listing disclosures, while AgiBot's latest revenue figures have been communicated by the company and reported through Chinese business media rather than an equivalent public prospectus. We can be highly confident that AgiBot is now one of the two largest businesses. Its claim to the number-one spot is convincing, but less independently transparent than Unitree's financial record.
If you want more recent data on this point, please see our latest humanoid robotics market report.
This chart, featured in our humanoid robotics market deck, illustrates yearly venture capital funding for humanoid robotics startups
How much money is Unitree actually making from humanoid robots?
Unitree made roughly RMB868 million from humanoid robots in 2025, turning humanoids into its largest business surprisingly quickly.
Unitree's disclosures show humanoids accounting for only 1.9% of company revenue in 2023. By 2025, they represented roughly half of RMB1.70 billion in total sales.
Unitree shipped more than 5,500 humanoids during 2025. Dividing RMB868 million by that volume gives a rough RMB158,000 of humanoid revenue per shipped unit. We should not read that as an exact average selling price because services, model mix and revenue-recognition timing can move the number, but it tells us something important about Unitree's model.
Unitree is reaching scale partly by selling relatively inexpensive humanoids in volume. Its G1 pushed the entry price far below many industrial humanoid systems, while the newer R1 pushed affordability further. That makes Unitree's economics quite different from companies trying to earn hundreds of thousands of dollars from each industrial deployment.
Is Unitree already profitable from robotics?
Yes. Unitree is currently the strongest example of a major humanoid robotics company combining large revenue with actual profit.
Unitree reported approximately RMB288 million of attributable net profit for 2025. Its gross margin was around 60%, unusually high for a young hardware manufacturer producing physical robots.
UBTECH shows how rare that is. UBTECH generated RMB2.00 billion of total company revenue in 2025, including RMB820.6 million from full-size humanoids, but remained deeply loss-making. Leju produced RMB258 million of sales and lost about RMB70 million. Agility's revenue was still tiny compared with the cost of developing and industrializing Digit.
AgiBot has not disclosed profit with anything close to Unitree's level of detail. Management has made clear that the company is still spending aggressively on growth.
So if “making the most money” means profit rather than revenue, Unitree deserves the clearest number-one answer today. We can verify both large-scale humanoid sales and a profitable company underneath them.
If you want more recent data on this point, please see our latest humanoid robotics market report.
This chart, featured in our humanoid robotics market deck, shows how Agility Robotics is capturing share in humanoid robotics
How fast is UBTECH's humanoid business growing now?
UBTECH's humanoid robot business is growing extraordinarily fast: first-half 2026 full-size humanoid revenue reached RMB590.3 million, more than fifteen times the level a year earlier.
UBTECH generated just RMB38.2 million from full-size embodied humanoids during the comparable first half of 2025. One year later, that segment supplied 46.5% of the company's RMB1.27 billion revenue.
Volume jumped just as sharply. UBTECH sold 921 full-size humanoids during the six months, compared with 45 a year earlier.
UBTECH also earns far more revenue per robot than Unitree. Its 2025 RMB820.6 million of full-size humanoid revenue came with roughly 1,079 Walker-series deliveries, which works out to around RMB760,000 per delivered robot before adjusting for services and product mix. Unitree's corresponding rough calculation is closer to RMB158,000.
That gap fits the companies' strategies. UBTECH concentrates more heavily on industrial deployments, enterprise solutions and customized projects. Unitree has built a broader, lower-priced volume business spanning research, development, education, entertainment and increasingly industry.
UBTECH's remaining problem is profit. Even with first-half revenue doubling company-wide and gross margin improving from 35.0% to 44.7%, the company still reported a large attributable loss. Its humanoid business is already commercially serious; its economics still need work.
| Company / period | Humanoid revenue | Approx. units | Rough revenue per unit |
|---|---|---|---|
| UBTECH 2025 | RMB820.6m | ~1,079 | ~RMB760k |
| Unitree 2025 | RMB868m | >5,500 | ~RMB158k |
| UBTECH H1 2026 | RMB590.3m | 921 | ~RMB641k |
Is China really dominating humanoid robot sales today?
Yes. China currently has an overwhelming lead in both humanoid robot shipments and the revenue we can actually verify.
Counterpoint estimated that AgiBot, Unitree and UBTECH together generated nearly 56% of global humanoid robot revenue in 2025. Other industry trackers also put Chinese manufacturers at the large majority of global unit shipments.
The difference becomes clearer when we compare individual companies. AgiBot and Unitree each shipped more than 5,000 robots in 2025. Leju, a much smaller Chinese competitor, sold 577 Kuavo full-size humanoids. UBTECH delivered roughly 1,079 Walker-series units.
Most famous American humanoid startups were still working with much smaller fleets. Agility's audited sales remained below $2 million. Figure had begun paid deployments but did not disclose anything approaching nine-figure robot revenue. Apptronik had major commercial agreements without comparable recognized revenue.
China's advantage today comes from commercialization as much as manufacturing capacity. The leading Chinese companies already have enough customers, production volume and price variety to turn thousands of robot deliveries into substantial sales.
If you want more recent data on this point, please see our latest humanoid robotics market report.
This chart, featured in our humanoid robotics market deck, illustrates yearly funding for humanoid robotics startups
Which smaller humanoid robotics startup is making real revenue?
Leju Robotics is the clearest company behind the top Chinese trio, with RMB258 million of 2025 revenue and almost RMB178 million coming from its Kuavo full-size humanoids.
Leju's Kuavo sales rose from only 32 units in 2024 to 577 in 2025. Revenue from the line increased to RMB177.8 million and represented 69.5% of the company.
The way Leju achieved that growth is useful. Its disclosed average Kuavo selling price fell from about RMB414,000 to RMB308,000, a 25.6% drop. Unit volume meanwhile increased about eighteenfold. The company was deliberately trading some price for much greater scale.
Leju's prospectus also shows what those robots are actually being used for. About 45% of Kuavo revenue came from data-collection applications, roughly one-third from research and education, and only about 4% from industrial manufacturing. Despite the public fascination with humanoids assembling cars, Leju's commercial business today is still mostly being built elsewhere.
The company remains loss-making, with a roughly RMB70 million 2025 net loss. Even so, Leju gives us a useful cutoff: a second-tier Chinese humanoid startup can already generate tens of millions of dollars in annual revenue without being anywhere near the scale of AgiBot, Unitree or UBTECH.
How much revenue are Figure, 1X, Apptronik and Agility actually making?
Western humanoid startups remain far behind the Chinese revenue leaders on the numbers we can verify today.
1X has built the largest clearly disclosed revenue base among this group. Norwegian company filings show sales revenue rising from roughly NOK6 million in 2024 to around NOK320 million in 2025, equivalent to roughly $30 million depending on the exchange rate used. The comparison needs care because that revenue is not purely from the new bipedal NEO. 1X previously commercialized its wheeled EVE robot, particularly for security applications.
Agility Robotics provides much cleaner humanoid-specific evidence and a much smaller number. Its audited 2025 net sales were approximately $1.8 million, including about $1.55 million from robot sales.
Figure has confirmed paying commercial deployments and has worked with BMW, but it still has not published a revenue figure remotely comparable with AgiBot, Unitree or UBTECH.
Apptronik has raised close to $1 billion and signed partnerships involving companies such as Mercedes-Benz and Jabil. We still lack sufficiently reliable Apollo revenue disclosure to rank it alongside the commercial leaders.
For now, funding has raced much further ahead of revenue in the American humanoid ecosystem.
| Company | Best disclosed revenue evidence | What we can conclude |
|---|---|---|
| 1X | ~NOK320m 2025 sales | Meaningful robotics revenue, but not all from NEO |
| Agility Robotics | ~$1.8m 2025 net sales | Humanoid commercialization still early |
| Figure AI | Current revenue undisclosed | Paying deployments exist; scale unclear |
| Apptronik | Current Apollo revenue undisclosed | Major partnerships; recognized sales unclear |
| Sanctuary AI | Current revenue undisclosed | Deployments exist; financial scale unclear |
This chart, featured in our humanoid robotics market deck, compares the main business model options for humanoid robot manufacturers
Does Agility Robotics' $300 million order book mean it is already a big business?
No. Agility Robotics has more than $300 million of announced orders but only about $1.8 million of 2025 net sales, so most of that commercial value still lies in the future.
Agility says those orders cover roughly 1,000 Digit v5 robots, with much of the business expected to run through multi-year robot-as-a-service contracts. Under an illustrative five-year model, one Digit could produce around $500,000 of cumulative revenue.
Those numbers make Agility interesting because customers are committing money before large-scale deployment. They also show why order books need to stay separate from recognized sales.
The more than $300 million headline is roughly 170 times Agility's 2025 revenue. Agility itself has warned that the order value depends on future contractual milestones and does not represent current-period revenue.
Agility could climb this ranking quickly once hundreds of Digit robots are actively generating recurring fees. As of now, it belongs in the “large future pipeline, small current revenue” category.
If you want more recent data on this point, please see our latest humanoid robotics market report.
Is Figure AI commercially as big as its valuation suggests?
No. Figure AI is valued like a future humanoid giant while its disclosed commercial revenue remains tiny compared with today's leaders.
Figure began generating commercial revenue after placing robots with BMW and has since continued developing its manufacturing strategy and general-purpose Helix intelligence stack. What Figure has never disclosed is anything resembling the RMB800 million-plus annual humanoid revenue already produced by Unitree and UBTECH.
That creates one of the widest valuation-to-revenue gaps in the sector. Figure's private valuation has climbed into the tens of billions of dollars while Chinese companies generating vastly more robot revenue have traded or raised capital at much lower valuations.
Investors are effectively paying for Figure's chance of becoming enormous rather than for the business it has already built. That can work if Figure eventually deploys robots by the tens or hundreds of thousands. It tells us very little about who makes the most money today.
This chart, featured in our humanoid robotics market deck, shows the revenue mix across customer segments in the humanoid robotics market
Are humanoid robots already making most of their money in factories?
No. Humanoid factory work gets most of the attention, but current revenue still comes heavily from data collection, research, entertainment and other non-production uses.
Counterpoint estimated that entertainment and performance applications generated about 26% of global humanoid sales revenue in 2025, while data-production applications supplied another 22%. Together, those categories represented almost half of the market.
Leju's disclosures give us an unusually detailed example. About 45% of its Kuavo full-size humanoid revenue came from data collection, while industrial manufacturing represented only 3.9%. Research and education supplied another large chunk.
Data-collection robots are being used to create physical-world training datasets: humans demonstrate manipulation tasks, robots record the movements and those recordings help train embodied-AI models. That creates revenue today even before robots are capable of replacing large numbers of human workers autonomously.
UBTECH is more exposed to industrial deployments, and Agility is specifically pursuing manufacturing and logistics with Digit. Those factory applications could eventually become much larger. The revenue mix today is still much broader and more experimental than the factory-demo headlines suggest.
How big is the humanoid robot market today?
The entire humanoid robot market is still only around the hundreds of millions of dollars in annual sales, despite individual startups being valued in the tens of billions.
Counterpoint put global 2025 humanoid robot sales revenue above $500 million. That makes the whole industry smaller than many single mid-sized technology companies.
AgiBot needed only roughly $150 million-scale humanoid sales to sit around the top of the global ranking. Unitree generated roughly $120 million from humanoids at typical 2025 exchange rates. UBTECH was in a similar range. Leju's Kuavo business was around $25 million.
Those numbers make the valuation gap hard to miss. Figure has commanded a valuation in the tens of billions. Apptronik has raised close to $1 billion. Agility has attracted a multibillion-dollar valuation despite less than $2 million of 2025 sales.
There is nothing inherently irrational about investing ahead of a new industry's revenue curve. The point for this ranking is simpler: most of today's humanoid valuations assume many years of future growth that have not yet arrived in the income statement.
This chart, featured in our humanoid robotics market deck, shows how factory humanoid robot technology has evolved over time
Which humanoid robotics startup is actually making the most money today?
AgiBot is currently the best answer for the private-startup revenue crown, while Unitree has the strongest proven combination of humanoid revenue and profitability.
AgiBot generated RMB1.05 billion of company revenue in 2025 and says it exceeded RMB1 billion again during the first quarter of 2026. That recent acceleration puts its current sales pace ahead of every other private humanoid startup we could verify.
Unitree remains the cleaner financial benchmark. It generated roughly RMB868 million from humanoids in 2025, RMB1.70 billion company-wide and approximately RMB288 million of attributable net profit. No other major independent humanoid player currently gives us such strong evidence of high volume, large revenue and positive earnings at the same time.
UBTECH sits close behind on humanoid sales. It generated RMB820.6 million from full-size humanoids in 2025 and another RMB590.3 million in the first half of 2026. Its continuing losses keep it behind Unitree when we look at the money the company actually keeps.
After those companies, revenue drops quickly. Leju has become a genuine second-tier commercial player. 1X has meaningful robotics sales, although its revenue is not purely bipedal humanoid revenue. Agility still had less than $2 million of annual sales despite a large order book. Figure, Apptronik and Sanctuary AI have yet to disclose revenue that would place them anywhere near the leaders.
So the companies actually making the most money from humanoid robotics today are mostly Chinese. AgiBot appears to lead private-startup revenue right now, and Unitree is the most financially proven business in the field. The Silicon Valley names attracting enormous valuations may eventually overtake them, but their present revenue does not support that ranking yet.
If you want more recent data on this point, please see our latest humanoid robotics market report.
OUR METHODOLOGY
This analysis asks which humanoid robotics startups are actually making the most money today. We compare realized revenue, profitability, shipment and deployment scale, recent growth, commercial mix and forward demand, while keeping funding and valuation separate from current commercial performance.
We use 2025 as the main common financial base because it gives the cleanest full-year comparison, then use 2026 quarterly and half-year figures where available to see whether the ranking is already changing. Humanoid-specific revenue carries more weight than broad company revenue; when a company also sells quadrupeds, wheeled robots or other robotics products, we say so rather than treating every dollar as humanoid revenue.
Orders and announced commercial agreements are treated as forward-demand evidence, not as current sales. The clearest example is Agility Robotics: its more than $300 million of announced Digit orders is commercially important, but its own SEC materials say that value is expected to be realized over time and is not a measure of current-period revenue.
We prioritized regulatory filings, exchange disclosures, audited or official financial results, company announcements containing concrete operating figures, and specialist industry research. Counterpoint Research is the main market-wide source for 2025 global humanoid revenue, AgiBot's estimated position and application mix; Shanghai Stock Exchange and Unitree listing materials provide the strongest financial detail for Unitree; HKEX filings provide UBTECH's 2025 and first-half 2026 results; and the Shenzhen Stock Exchange prospectus provides Leju Robotics' revenue, Kuavo sales, pricing and application mix.
Other key sources used include Counterpoint Research on the global humanoid market, Yicai on AgiBot's 2025 revenue, AgiBot's 10,000-unit production milestone, AgiBot's 15,000-unit production milestone, Unitree's Shanghai Stock Exchange prospectus, the Shanghai Stock Exchange summary of Unitree's 2025 financials and shipments, UBTECH's 2025 annual results, UBTECH's first-half 2026 results, Leju Robotics' prospectus, Norway's Brønnøysund Register Centre for 1X filings, 1X on EVE's commercial history, Agility Robotics' SEC transaction announcement, Agility's SEC investor presentation, Figure AI's Series C and valuation announcement, BMW Group on Figure's factory deployment, Apptronik on its funding and commercial relationships, Apptronik and Jabil on their manufacturing partnership, and Sanctuary AI on its commercial humanoid deployment.
In our humanoid robotics market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
- Which humanoid robotics companies are actually making money?
- Which humanoid robotics company has the biggest order book?
