What are the most valuable humanoid robotics startups?

Top startups in the humanoid robotics market ranked by valuation

Here is an updated table that ranks the top startups in the humanoid robotics market based on their latest reported or estimated valuations.

If you need to dig deeper and get more detailed data, please check our report covering the humanoid robotics market.

1. Total funding is a minimum: some rounds have no public amount.

Which humanoid robotics startups are worth the most today?

Figure AI is by far the most valuable humanoid robotics startup right now, with a valuation of $39.0B. AgiBot follows at $5.5B–$7.5B, then Apptronik at $5.3B, 1X Technologies at $4.0B–$6.0B, and NEURA Robotics at about $4.4B. Figure's lead is striking because the industry is still at an early stage of commercial adoption.

Figure announced more than $1 billion in Series C commitments in September 2025 at a $39 billion post-money valuation. Investors included NVIDIA, Brookfield, Intel Capital, and Salesforce, alongside lead investor Parkway Venture Capital. The company is betting that its Helix AI system and humanoid manufacturing operations can eventually serve several huge markets, from automotive production to household assistance.

The other leading companies have different strengths. AgiBot is already shipping robots in substantial quantities. Apptronik is developing its Apollo humanoid for industrial customers, supported by strategic investors including Google and Mercedes-Benz. 1X Technologies is aiming at homes with its NEO robot, while NEURA Robotics is building an ecosystem of cognitive robots and shared AI capabilities.

What's interesting is how much investors currently value the possibility of building a general-purpose robot platform. Figure has an enormous financial lead, although AgiBot's shipment volumes and Agility Robotics' operational track record show that the companies making the most progress commercially are not necessarily those with the highest valuations.

Why are humanoid robotics startup valuations growing so fast?

Humanoid robotics startups are attracting much larger investments because AI has made general-purpose robots more believable, and manufacturers are starting to put them to work. Investors are financing the possibility of a large new industry, years before its full commercial potential can be measured.

Goldman Sachs made a major change to its forecasts in February 2024. It increased its estimate for the global humanoid robotics market in 2035 from $6 billion to $38 billion and raised projected annual shipments to 1.4 million robots. Better AI capabilities and an estimated 40% decline in manufacturing costs were important reasons for that revision.

The financing that followed shows how quickly expectations changed. Apptronik announced a $415 million Series A in 2025 and extended it by another $520 million in February 2026. NEURA Robotics announced a Series C financing of up to $1.4 billion in June 2026. Figure's September 2025 round crossed $1 billion in commitments.

Those are enormous amounts for companies still building their first large-scale commercial operations. Investors expect the eventual opportunity to extend beyond selling machines. Robot intelligence, software updates, fleet management, and specialized applications could generate additional revenue over a robot's working life.

The gap between that future opportunity and the market today is huge. Counterpoint Research estimated that global humanoid shipments exceeded 22,000 in the first half of 2026, while the International Federation of Robotics recorded more than 600,000 new installations of conventional industrial robots during 2025. Humanoids are growing much faster from a small base, but they still represent a tiny fraction of the broader industrial robotics business.

How many humanoid robotics startups are worth more than $1 billion?

Our ranking contains 27 humanoid robotics startups whose entire valuation ranges are at least $1 billion, plus four more whose ranges cross that threshold. That gives us 31 companies with a plausible billion-dollar valuation, although only 27 clear the threshold throughout their estimated ranges.

Among the largest are Figure AI, AgiBot, Apptronik, 1X Technologies, and NEURA Robotics. Further down, startups such as X Square Robot, Astribot, RobotEra, and AI² Robotics also have valuations above $1 billion.

The distinction becomes more important around the threshold. Leju Robotics is valued at $850M–$1.4B, Anyverse Dynamics at $800M–$1.2B, Moqi Intelligent Technology at $800M–$1.2B, and Futuring Robot at $700M–$1.2B. All four could qualify as unicorns, but their valuation ranges also extend below $1 billion. Foundation Robotics, by contrast, reached $3.4B with its Series A-2.

There is also a difference between a valuation established through a completed financing transaction and one estimated from a company's progress and comparable businesses. Our methodology accounts for that distinction.

We therefore count startups with a valuation of at least $1 billion across their entire range separately from those that may have reached that level. The broader figure of 31 illustrates the scale of investor interest, while 27 represents the companies that clear the threshold even at the lower end of their ranges.

Is Figure AI really worth six times more than other humanoid robotics startups?

Figure AI's $39.0B valuation is about six times the central value of AgiBot's $5.5B–$7.5B range ($6.5B) and more than seven times Apptronik's $5.3B. Its premium reflects an unusually ambitious bet on general-purpose AI and future commercial scale, rather than a demonstrated sixfold advantage in current business activity.

Figure has made genuine technical and operational progress. In November 2025, the company disclosed that its Figure 02 robots had spent more than 1,250 hours operating at BMW's Spartanburg plant, loading over 90,000 parts and contributing to the production of more than 30,000 BMW X3 vehicles. That works out to roughly 72 parts per reported operating hour.

The company then returned to BMW with Figure 03 in June 2026, demonstrating a more complex logistics task involving part sequencing and whole-body movement. In September, Figure also retired most of its Figure 02 fleet as it moved toward newer generations. That unusually fast hardware replacement cycle shows how quickly the technology is progressing, but also how much development work remains.

Meanwhile, AgiBot achieved approximately 9,700 shipments in the first half of 2026, according to Counterpoint Research. Figure has not published a comparable company-wide shipment figure.

So Figure's enormous valuation premium remains a bet on what its technology could eventually become. The BMW work and Helix demonstrations give that bet substance, but meaningful revenue, production volumes, and operating margins will ultimately determine whether Figure deserves such a large premium over its competitors.

Are Chinese humanoid robotics startups beating American companies?

Chinese humanoid robotics startups are currently winning the production-volume race, while American companies still command some of the industry's largest private valuations. The two countries are strong in different areas, and shipment volume is now an especially important Chinese advantage.

Counterpoint Research reported that AgiBot shipped approximately 9,700 humanoids in the first half of 2026, representing more than 43% of worldwide shipments. Galbot shipped more than 1,100, while Leju Robotics delivered roughly 650. Chinese manufacturers also benefit from domestic suppliers of motors, sensors, batteries, actuators, and precision components.

That manufacturing advantage extends well beyond humanoids. The International Federation of Robotics reported that China installed 354,000 industrial robots in 2025, accounting for 59% of global installations. Its enormous existing automation industry gives newer humanoid developers access to customers, suppliers, engineers, and manufacturing expertise.

American companies have other strengths. Figure has invested heavily in proprietary robot intelligence and manufacturing capacity. Agility Robotics has established commercial contracts with industrial customers, including GXO. Apptronik is developing Apollo with backing from major technology and manufacturing investors.

China has the lead in early humanoid manufacturing scale. The United States remains highly competitive in robot intelligence, enterprise relationships, and financing. The bigger question now is whether Chinese companies can turn their shipment advantage into higher profits.

Which humanoid robotics startups are selling the most robots?

AgiBot is the clearest shipment leader among the companies in our valuation ranking, but the latest industry data also shows that many humanoids are being sold for research, entertainment, and data collection rather than factory labor.

Counterpoint Research estimated more than 22,000 global humanoid shipments in the first half of 2026, up nearly 300% from the previous year. AgiBot accounted for approximately 9,700 units, followed within our ranking by Galbot with more than 1,100 and Leju Robotics with around 650.

One particularly interesting finding concerns what buyers are doing with those robots. Entertainment and performance applications, together with data production and research, accounted for more than 60% of shipments. Intelligent manufacturing represented 13%, and warehousing and logistics another 5%.

In other words, fewer than one in five shipped humanoids were going into those two industrial categories. The commercial market is growing quickly, but most of its early volume is still outside routine factory and warehouse labor.

Counterpoint expects more than 50,000 global shipments for the full year. Reaching that forecast would mean annual volume more than doubling the first-half total, although the composition of those shipments will be just as revealing as the final number.

Shipment indicator	Reported result	What we can conclude
Worldwide H1 2026 shipments	More than 22,000	Rapid growth from an early base
Annual shipment growth	Nearly 300%	Production is accelerating
AgiBot shipments	Approximately 9,700	Largest volume among ranked startups
AgiBot global share	More than 43%	Substantial supplier concentration
Galbot shipments	More than 1,100	Meaningful early production
Leju Robotics shipments	Approximately 650	Smaller-scale commercial output
Top five global suppliers	86% of shipments	A concentrated industry
Entertainment and research	More than 60% of shipments	Major early sources of demand
Intelligent manufacturing	13% of shipments	Factory use remains limited
Warehousing and logistics	5% of shipments	Logistics adoption is still early
Full-year forecast	More than 50,000	Strong further growth expected

Are humanoid robots actually doing useful work in factories now?

Humanoid robots are already doing real industrial work, especially moving materials, loading parts, and handling repetitive warehouse tasks. The strongest deployments are still quite specialized, but several have progressed well beyond short demonstrations.

Figure's work at BMW is one example. Its Figure 02 robots handled sheet-metal parts on an active assembly line, with reported 10-hour shifts on weekdays. The company also documented specific hardware lessons, including forearm reliability problems that influenced the redesigned wrist electronics of Figure 03.

Agility Robotics provides another example from logistics. Its Digit robots have been working at GXO since 2024, moving totes and placing them onto conveyors. By October 2026, the company had reported more than 100,000 totes moved at GXO and over 65,000 operating hours across customer programs and commercial operations.

Apptronik is pursuing industrial applications with Apollo, while NEURA Robotics is developing cognitive systems designed to work in industrial environments. Their commercial progress is promising, although public operating figures remain less detailed.

The useful tasks today are relatively narrow: moving containers, handling components, and performing repeated sequences. That is enough to make humanoids commercially interesting, particularly in facilities built around human workstations. Handling a much wider variety of tasks without reprogramming, though, is still difficult.

Is Agility Robotics closer to making money from humanoids than Figure AI?

Agility Robotics has stronger public evidence of repeat commercial customers, although Figure AI is also proving that its robots can work inside factories. Agility's advantage is the continuity of its logistics contracts and the expanding number of industrial customers willing to test or deploy Digit.

Agility's relationship with GXO began as a pilot in late 2023 and became a multiyear Robots-as-a-Service agreement in June 2024. In October 2026, the two companies announced plans to deploy Digit at two additional GXO facilities in North Carolina and Tennessee beginning in 2027.

The company has also disclosed a larger set of commercial relationships. Toyota Motor Manufacturing Canada, Schaeffler, Mercado Libre, and AGCO are among the customers or partners involved in Digit programs. Agility reported more than $300 million of multiyear contracted orders for Digit 5 when announcing its planned public-market transaction.

Figure's work at BMW has been substantial, and its newer Figure 03 system is targeting more complex tasks. Yet Figure has not released equivalent company-wide information about contracted orders or the economics of its deployments.

Agility's proposed transaction with Churchill Capital Corp XI included a $2.5 billion pre-money equity value and more than $620 million in expected gross proceeds. These terms offer a useful financing benchmark, although the planned listing and actual commercial revenue are separate milestones.

Agility is one of the strongest commercial-execution stories in humanoid robotics today. Figure has broader ambitions and a much higher valuation; Agility currently has more concrete proof of repeat customers and contracted demand.

Why are investors putting nearly $1 billion into Apptronik?

Investors are putting enormous sums into Apptronik because its Apollo robot is being developed for large industrial customers, and scaling production will require much more money than building prototypes. The company has attracted an unusually strong collection of manufacturing and technology investors.

Apptronik announced a $520 million extension of its Series A in February 2026, bringing that round to more than $935 million and its total funding to more than $950M. Existing investors included Google, Mercedes-Benz, B Capital, and PEAK6, while newer participants included John Deere, AT&T Ventures, and Qatar Investment Authority.

The composition of the investor group is important. Mercedes-Benz and John Deere operate in industries where repetitive physical work and factory automation are significant costs. Their involvement gives Apptronik potential opportunities to develop robots around real operational requirements.

NEURA Robotics has attracted similar interest through a different strategy. Its June 2026 Series C, described as reaching up to $1.4 billion, included backing associated with major technology and industrial groups. NEURA intends to expand its Neuraverse platform and establish training environments for cognitive robots.

Both companies are trying to finance the transition from development into industrial-scale production. Apptronik's funding gives it resources to pursue that transition, but how much customers are ultimately willing to pay for Apollo will be a more useful test than the amount of capital raised.

Can 1X Technologies really sell humanoid robots for people's homes?

1X Technologies has made household humanoid robots available to order, but its NEO product still needs human assistance for some complex tasks. The commercial offer is real; fully autonomous household service remains a work in progress.

NEO is offered through a $20,000 early-access ownership package or a $499 monthly subscription scheduled for a later rollout. The company announced initial US deliveries for 2026, with international expansion planned from 2027.

The more revealing detail is NEO's Expert Mode. Customers can schedule a remote operator to help the robot complete unfamiliar tasks and learn new abilities. That could make the product useful earlier than a completely independent household robot, but it also means some tasks still depend on human work behind the scenes.

The monthly subscription costs $5,988 per year. For an ordinary household, that's a meaningful expense, particularly if the robot's useful capabilities remain limited to a relatively small number of chores.

Figure is pursuing household autonomy too. Its Helix 02 system demonstrated a four-minute autonomous dishwasher task in January 2026, and subsequent work expanded its domestic-task demonstrations.

These are genuine advances, but 1X still has to prove that NEO can consistently handle enough everyday chores to make the cost worthwhile for a large number of households. That's the main commercial hurdle for home humanoids.

Will humanoid robots be more successful in homes or factories?

Factories are currently the much stronger market for humanoid robots because customers have repetitive work, established automation budgets, and clear ways to measure productivity. Household robotics could eventually become enormous, but the technology needs to become much more reliable and versatile.

Industrial demand is already visible. BMW has tested Figure robots in automotive production, GXO is expanding its work with Agility, and Apptronik has relationships with manufacturing groups. These customers can evaluate how many objects a robot moves, how often it fails, and whether deployment reduces operating costs.

Homes are harder. A household robot might need to fold laundry, clean kitchens, handle fragile dishes, recognize unfamiliar objects, and work safely around people. Each task adds new technical requirements.

Goldman Sachs' February 2024 forecasts anticipated more than 250,000 annual humanoid shipments by 2030, almost entirely for industrial applications. It expected consumer adoption to grow later, once hardware costs and technical limitations improved.

Even industrial adoption has plenty of room to grow. According to the International Federation of Robotics, factories worldwide installed approximately 603,000 conventional industrial robots in 2025 and had more than five million operating. Humanoids are entering a huge automation market, but they first have to prove that their flexibility justifies choosing them over conventional machines.

For now, factories offer the clearest route to customers who will pay repeatedly for useful work. Household robots have a much wider range of possible tasks, and that's precisely what makes them so difficult to build.

Are humanoid robots getting cheap enough to replace workers?

Humanoid robot prices are falling, but that doesn't automatically make them cheaper than human workers. Purchase prices are improving much faster than the evidence about reliability, supervision costs, and productive operating hours.

Goldman Sachs estimated in February 2024 that manufacturing costs had fallen from roughly $50,000–$250,000 per robot to $30,000–$150,000, depending on the type of machine. More affordable components, better designs, and improved manufacturing methods were behind much of the decline.

NEO's $20,000 early-access price shows that some household humanoid hardware is already approaching the cost of an expensive consumer product. Industrial machines with greater payload capacity, precision, and reliability requirements can cost substantially more.

Consider a hypothetical $30,000 robot used over five years. Its hardware cost would be $6,000 annually before maintenance, software, power, financing, or operator support. At 2,000 productive hours per year, that amounts to $3 per hour in hardware depreciation. At only 500 hours, it becomes $12 per productive hour.

That's a fourfold difference for exactly the same machine. A relatively inexpensive robot that repeatedly fails or needs remote supervision may deliver worse economics than a costly machine that runs reliably throughout a shift.

Cost per completed task deserves more attention than advertised robot prices. It's the measure that will ultimately determine whether industrial customers buy hundreds of humanoids rather than a handful.

What makes some humanoid robotics startups more attractive to factories?

Factories are most interested in humanoid robotics startups that can handle useful tasks reliably and fit into existing operations without expensive changes. A robot with impressive movement capabilities has little commercial value if it cannot meet the customer's required speed, accuracy, and uptime.

Different startups are targeting different parts of that problem. Figure AI has demonstrated automotive part handling and logistics sequencing. Agility Robotics focuses heavily on warehouse workflows. Apptronik is developing Apollo for industrial applications, while NEURA Robotics combines humanoids with a wider cognitive robotics platform.

Other companies are more specialized. Persona AI is targeting industrial welding. Galbot, RobotEra, and AI² Robotics concentrate on embodied intelligence and manipulation, which could eventually support a wider range of industrial operations.

The level of commercial proof varies considerably. An operating deployment tells us more than a pilot announcement, and a repeat order says more about customer satisfaction than a single trial.

Safety matters too. ISO 10218-2:2025 sets requirements for integrating industrial robot applications and cells. Humanoids working near people must be assessed for the actual hazards of their tasks and environments, with additional requirements where the standard does not fully cover a mobile configuration.

Startup	Industrial focus	Evidence of progress
Figure AI	Automotive production and logistics	BMW deployment and subsequent Figure 03 trials
Agility Robotics	Warehouse and factory material handling	GXO agreement, expansion and operating records
Apptronik	Manufacturing automation	Apollo development and industrial relationships
NEURA Robotics	Cognitive industrial robotics	Large Series C and manufacturing expansion plans
AgiBot	Industrial and service robots	High reported shipment volume
Galbot	Embodied AI and manipulation	More than 1,100 estimated H1 2026 shipments
Leju Robotics	Industrial humanoids	Approximately 650 estimated H1 2026 shipments
Persona AI	Industrial welding	Specialized humanoid welding development

Are humanoid robotics startups becoming more like AI companies?

The most ambitious humanoid robotics startups now spend enormous effort developing robot intelligence because reusable AI could become more valuable than the hardware itself. A robot that learns additional tasks through software has a much larger market than one designed for a single operation.

Figure's Helix system is a clear example. In January 2026, Figure introduced Helix 02, combining high-level reasoning, visual understanding, and whole-body control. Its demonstrations included a continuous dishwasher task, while the technical system integrated cameras, tactile sensors, and motion control.

The company later demonstrated Helix 2.5 generalizing across 30 home environments. That work is significant because robots need to cope with unfamiliar objects and layouts without being individually programmed for every situation.

NEURA Robotics is developing Neuraverse as a shared intelligence ecosystem, while Galbot, X Square Robot, Spirit AI, and GigaAI are also focusing on embodied AI and robot learning. Their goal is to make useful robotic behavior more transferable across machines, tasks, and environments.

Dexterity is becoming especially important. Sharpa and Astribot are developing systems centered on sophisticated manipulation, while Figure has demonstrated tactile-assisted tasks involving small objects and controlled force.

The commercial attraction is straightforward: a developer that can improve thousands of deployed robots through software updates has an advantage over one that must redesign hardware for every new task. However, robot companies will still need to manufacture machines, repair them, and support installations. Software can improve the economics, but it won't make those physical costs disappear.

Could cheaper Chinese humanoid robots hurt American startups?

Cheaper Chinese humanoid robots could put real pressure on American manufacturers, particularly when customers can get similar performance for much less money. The biggest risk is that hardware becomes increasingly interchangeable before startups establish profitable software and service businesses.

China already has a strong manufacturing base for robot components. TrendForce forecast in April 2026 that Chinese humanoid output could grow by approximately 94% during the year, supported by production expansion and a clearer focus on commercial applications.

AgiBot's shipment lead suggests that some Chinese manufacturers are beginning to benefit from scale. Larger production runs can help suppliers improve component sourcing, manufacturing efficiency, and product reliability.

American developers may be able to defend higher prices through better autonomy, specialized capabilities, or stronger customer support. Figure's Helix technology, Apptronik's industrial partnerships, and Agility's commercial logistics deployments offer different ways to compete.

The threat becomes more serious as hardware improves across the industry. A factory comparing two robots will eventually care more about useful operating hours and cost per task than where the technology was developed.

Price competition will become increasingly important for standardized humanoid platforms. Companies offering proprietary intelligence, reliable dexterity, or better performance in demanding industrial environments should have more room to protect their margins.

Are humanoid robotics startups profitable yet?

There is still too little reliable financial disclosure to establish that the leading private humanoid robotics startups are consistently profitable. Several have meaningful customers, financing, and shipment volumes, but those figures cannot tell us whether they are making money.

The scale of investment is considerable. Figure AI has raised at least $1.85B, Galbot at least $982M, and Apptronik $950M+, according to the financing records reflected in our ranking. NEURA Robotics announced a Series C of up to $1.4 billion, but our funding column, which leaves out commitments not yet paid, counts only $215M+ for it.

Much of that capital supports work that must happen before substantial profits are possible. Robot developers need to design actuators and hands, train AI models, build production lines, test safety systems, and maintain machines operating at customer sites.

The current shipment mix also limits what we can infer about earnings. A robot sold for research or entertainment may have a different margin and support requirement from one used continuously in a warehouse. Even large shipments don't reveal how much money remains after production and operating costs.

Agility Robotics offers a clearer view of contracted demand, with more than $300 million of multiyear Digit 5 orders disclosed alongside its planned public-market transaction. But contracted orders represent future business commitments, not current revenue or profit.

Humanoid robotics has become a commercial industry. Consistent profitability across its leading startups is another matter, and one that could take years to establish.

Are humanoid robotics startup valuations too high right now?

Humanoid robotics startup valuations are ahead of what today's commercial market can clearly support, although the technology has made enough progress to justify serious long-term investment. The biggest private valuations depend on an industry becoming dramatically larger over the coming decade.

The evidence for growth is convincing. Counterpoint recorded nearly 300% year-over-year shipment growth in the first half of 2026. AgiBot is shipping thousands of robots. Figure has completed a substantial BMW deployment, and Agility is expanding a commercial logistics relationship that began in 2024.

The financial scale is still much smaller than the valuations might suggest. Counterpoint expects global shipments to exceed 50,000 units in 2026. At a hypothetical average selling price of $30,000, exactly 50,000 units would generate $1.5 billion in annual hardware sales across the entire global industry. That is an illustrative calculation rather than a reported revenue figure, but it puts the current size of the opportunity into perspective.

Some valuations may eventually prove reasonable if deployments grow rapidly, autonomy improves, and customers begin ordering large robot fleets. Figure's $39.0B valuation is particularly dependent on that outcome. AgiBot, Apptronik, 1X Technologies, and NEURA Robotics also have ambitious growth expectations built into their multibillion-dollar valuations.

There are three major risks: robots may take longer than expected to perform useful tasks reliably, manufacturing costs may remain high, and customers may find cheaper conventional automation solutions. A slowdown in financing could make those problems worse for companies still spending heavily on development.

The next winners will be the companies that turn early deployments into repeat orders, sustained operating hours, and eventually profitable fleets. Investors are paying a great deal today for that possibility, and only some startups will deliver it.