What are the top startups in the humanoid robotics market?
In our humanoid robotics market deck, you will find everything you need to understand the market
SUMMARY
The top startups in the humanoid robotics market are Figure AI, Agility Robotics, Unitree, AgiBot, Apptronik, Neura Robotics, 1X, Skild AI and Physical Intelligence.
The market is not led by one type of company. Figure and Agility lead on real-world proof, Unitree and AgiBot lead on manufacturing and cost, while Skild AI and Physical Intelligence matter because the robot-brain layer may capture a large share of the value.
The strongest signal in humanoid robotics is no longer the demo video. The better test is whether a company can publish operating metrics: runtime hours, parts moved, totes moved, revenue, production milestones or customer deployments.
Figure AI has the strongest Western all-around story because it combines capital, a BMW factory deployment and a newer logistics agreement. But that also makes it one of the most demanding valuation stories in the market.
Agility Robotics looks less broad than Figure, but its logistics evidence is cleaner. Moving more than 100,000 totes in live warehouse work is narrow, but it is easier to judge than vague claims about general-purpose autonomy.
Unitree is the company most likely to change the market from below. Its low-cost humanoids and public revenue signals make humanoid robotics feel more like an accessible developer platform than a closed enterprise deployment category.
AgiBot is the most under-discussed Chinese challenger. Its reported 10,000-unit milestone needs caution, but even cautious readers should treat it as a serious manufacturing-scale signal.
Neura Robotics changed status because of capital and industrial backing, not because it has already proven large-scale deployment. The company now has the resources and channel access to matter, but its proof still needs to catch up.
1X owns the home-humanoid narrative because it is the only major player making a direct consumer push. That makes it exciting, but also unusually exposed to safety, reliability, privacy and support challenges.
Skild AI and Physical Intelligence show why the humanoid market should not be ranked only by robot bodies. Investors are clearly betting that reusable robot intelligence could become a platform layer across many hardware forms.
The most useful ranking is therefore not a single leaderboard. The better answer is category-based: real-world proof, manufacturing scale, affordability, industrial customers, capital momentum, robot intelligence, China exposure, home robotics and valuation discipline.
The final conclusion is that Figure AI, Agility Robotics, Unitree and AgiBot have the strongest operating signals today, while Apptronik, Neura Robotics, 1X, Skild AI and Physical Intelligence belong in the top group because they repeatedly appear across capital, customer, product and AI-stack evidence.
This market map, featured in our humanoid robotics market deck, highlights top companies and startups in the humanoid robotics market
Which humanoid startups have proof outside the demo video?
Figure AI and Agility Robotics lead this category, with Apptronik behind them and Reflex Robotics still too early to rank as a proven leader.
Figure is currently the strongest Western deployment story because its BMW pilot moved from “cool robot in a factory” to measured factory work. In November 2025, Figure said two Figure 02 robots completed an 11-month BMW Spartanburg deployment, worked 10-hour shifts from Monday to Friday, logged more than 1,250 runtime hours, loaded more than 90,000 parts, and supported more than 30,000 BMW X3 vehicles. The important comparison is not against Tesla Optimus or Unitree videos but against other private humanoid startups that still rarely publish task-volume data. On that standard, Figure is ahead.
Agility Robotics is the other company with unusually concrete proof. In November 2025, Digit passed more than 100,000 totes moved in live commercial work at GXO. That is a narrower use case than Figure’s general-purpose ambition, but the metric is cleaner. A tote either moved or it did not. For logistics, Agility has the most legible throughput signal in the market.
Apptronik belongs in the second tier here. It has strong enterprise credibility, with Google, Mercedes-Benz and GXO-related signals around Apollo, but the public evidence is still more about pilots, investors and partnerships than repeated task volume. Reflex Robotics is even earlier: GXO pilot visibility makes it worth watching, but not enough to put it beside Figure or Agility yet.
If you want more recent data on this point, please see our latest humanoid robotics market report.
Which startups look closest to becoming real robot manufacturers?
Unitree and AgiBot look strongest on manufacturing scale, while Neura Robotics and Apptronik are now trying to buy their way into that conversation with fresh capital.
Unitree is the cleanest manufacturing story because it has a public-market filing, real revenue and unusually low prices. Its IPO review materials show revenue rising from RMB 159 million in 2023 to RMB 393 million in 2024, then to nearly RMB 1.7 billion in 2025. It also expects roughly RMB 1.05 billion to RMB 1.13 billion in revenue for the first half of 2026. That puts Unitree in a different class from startups that only have lab prototypes and funding announcements. It is already behaving like a hardware company with measurable sales.
AgiBot may be the most aggressive unit-scale challenger. In March 2026, the company said it had reached its 10,000th humanoid robot, after going from 5,000 to 10,000 units in about three months. We should treat company-reported shipment milestones carefully, but the comparison is still striking: many U.S. startups discuss future fleets, while AgiBot is already pushing a volume narrative.
Neura Robotics is the fresh European shock. In June 2026, it announced up to $1.4 billion in funding and talked about scaling toward millions of robots by 2030. That is not delivered-unit proof, but it gives Neura the capital base to attempt industrial-scale production. Compared with Unitree and AgiBot, Neura is less proven on units today. Compared with almost every European robotics startup, it is suddenly in another weight class.
Apptronik sits between these groups. It has almost $1 billion in total capital after its February 2026 Series A extension, and the money is explicitly tied to Apollo scale-up. The gap is that Apptronik has not yet shown the same public revenue or unit-production proof as Unitree, nor the same claimed shipment scale as AgiBot.
As this chart shows, and as featured in our humanoid robotics market deck, search interest in where to buy robots has been rising steadily
Which startups are winning because their robots are actually affordable?
Unitree is far ahead on price. Fourier Intelligence is relevant for labs and institutions, but it is not changing the cost curve the way Unitree is.
Unitree’s advantage is simple: its humanoids are priced like developer hardware, not like factory capex. The R1 is listed from about $4,900, and the G1 starts around $13,500. That makes Unitree extremely hard to compare with Figure, Agility or Apptronik, because those companies are selling deployment outcomes rather than broadly accessible robot bodies. Still, price changes the market. A sub-$20,000 humanoid can reach developers, universities, content creators and smaller labs in a way a six-figure robot cannot.
Fourier Intelligence is the next credible name in this category, but for a different buyer. The GR-2 is around $125,000 in reseller listings and is aimed at research, rehab and institutional pilots. That is much more accessible than bespoke enterprise humanoid deployments, but it does not create the same bottom-up experimentation flywheel as Unitree.
The strategic takeaway is that Unitree is making humanoids feel like a platform. Most Western leaders still feel like enterprise robotics vendors. That difference matters because developer adoption can create weird, fast, unpredictable learning loops.
If you want more recent data on this point, please see our latest humanoid robotics market report.
Which startups are getting the best industrial customers?
Figure AI, Agility Robotics and Apptronik form the leading U.S. industrial pack, with Neura Robotics now entering through European industrial backers.
Figure is first because BMW gives it a high-quality automotive reference, and Catalyst Brands adds a fresher logistics angle. The BMW pilot proves the company can run a defined factory task over time. The May 2026 Catalyst Brands agreement matters because it tests whether Figure can move beyond one automaker into distribution and logistics. Compared with Agility, Figure’s wedge looks broader. Compared with Apptronik, its proof is more public.
Agility is narrower but more operationally mature in logistics. GXO’s tote-moving deployment is the kind of boring warehouse work that can become repeatable revenue if reliability holds. The comparison with Figure is useful: Figure is chasing broader general-purpose credibility; Agility is turning one painful logistics motion into a measurable automation product.
Apptronik has perhaps the strongest “enterprise likes this company” signal after Figure. Google, Mercedes-Benz, John Deere and other industrial names around the February 2026 financing make Apollo look like a serious enterprise platform. The reason we rank it below Figure and Agility today is evidence quality. Investors and partners are strong signals, but throughput beats logos.
Neura is the non-U.S. name to watch here. Its June 2026 round included Bosch, Schaeffler, Amazon, Nvidia, Qualcomm and the European Investment Bank. That investor mix is unusually industrial. We would not say Neura has beaten Figure or Agility on deployment evidence, but it may now have the best European channel into manufacturing and logistics customers.
This chart, featured in our humanoid robotics market deck, illustrates yearly venture capital funding for humanoid robotics startups
Which startups are becoming hot because of capital, not deployment?
Figure AI, Skild AI, Neura Robotics, Apptronik and Galbot are the clearest capital winners. But they are not all winning the same way.
Figure remains the valuation outlier. In September 2025, it announced more than $1 billion in Series C committed capital at a $39 billion post-money valuation. That is far above Apptronik’s reported $5 billion valuation and Neura’s reported $7 billion valuation. The market is effectively pricing Figure as the default private humanoid leader, not just one promising robotics startup.
Skild AI is the biggest reminder that the humanoid race is not only about bodies. In January 2026, it raised close to $1.4 billion at a valuation above $14 billion for its robotics foundation model. That puts Skild above most hardware startups by valuation, even though it does not sell a humanoid body. Investors are saying the software layer could capture more value than the robot shell.
Neura is the freshest capital story. Its June 2026 raise of up to $1.4 billion at a reported $7 billion valuation instantly made it Europe’s most serious full-stack humanoid contender. Compared with Figure, it is still less proven on public deployment metrics. Compared with Europe’s previous position in humanoids, it is a step-change.
Apptronik’s $935 million-plus Series A is impressive because it comes earlier in the company’s financing stack. A Series A of that size says investors are not just funding R&D; they are funding the attempt to industrialize Apollo. Galbot’s March 2026 RMB 2.5 billion round is smaller in dollar terms, but very important inside China because it combined state-linked capital, embodied-AI ambition and reported orders from major industrial groups.
If you want more recent data on this point, please see our latest humanoid robotics market report.
Which startups are winning the robot-brain race?
Skild AI and Physical Intelligence lead the independent software layer, while Figure AI and 1X are the most interesting body-plus-brain companies.
Skild AI is currently the most highly valued pure robot-intelligence startup. Its January 2026 round valued it above $14 billion, roughly 2.5 times Physical Intelligence’s reported $5.6 billion valuation from late 2025. That does not prove Skild has better technology, but it does show where investors see the biggest platform upside: a model that can transfer across robot bodies.
Physical Intelligence is the strongest counterweight. Its $600 million Series B in late 2025 was smaller than Skild’s round, but still huge for a company focused on robot foundation models. The difference is positioning. Skild talks heavily about omni-bodied intelligence; Physical Intelligence has become the reference name for general robot action models. Both could become suppliers to humanoid companies rather than humanoid winners themselves.
Figure is important because it is trying to integrate the body, deployment data and AI stack inside one company. If the BMW and logistics deployments generate useful data, Figure’s advantage could compound. It would not just have robots in the field but robots actually creating training data in commercially relevant settings.
1X is interesting for a different reason. Its home-humanoid bet forces the company to solve messy human environments, and its world-model work points in that direction. The risk is that home data is harder to collect safely and consistently than factory data. The upside is that if 1X cracks it, the learning environment is much larger than a warehouse.
This chart, featured in our humanoid robotics market deck, shows how Agility Robotics is capturing share in humanoid robotics
Which startups are really leading China’s humanoid push?
Unitree and AgiBot are the two strongest China leaders today, with Galbot rising fast and Fourier Intelligence occupying a more institutional lane.
Unitree leads because it combines three things most humanoid startups do not have at the same time: low price, revenue growth and IPO visibility. The RMB 1.7 billion 2025 revenue figure gives it a financial signal that venture-backed Western startups rarely disclose. Its cheap humanoids also make it more likely to spread quickly through developers and research buyers.
AgiBot looks like the most serious Chinese volume challenger. Its March 2026 10,000-unit milestone is hard to ignore, even if we keep some caution around company-reported numbers. Compared with Unitree, AgiBot’s public financial picture is less clear. Compared with Western startups, its shipment narrative is much more aggressive.
Galbot is not yet at Unitree or AgiBot’s scale, but it may be the most interesting “state-backed embodied AI” startup in China right now. Its March 2026 RMB 2.5 billion raise, prior late-2025 financing, reported RMB 20 billion-plus valuation and claimed orders from industrial names such as CATL, Bosch, Toyota, BAIC and SAIC give it a strong commercialization story. The missing piece is verified deployment volume.
Fourier Intelligence is more specialized. Its GR line has healthcare, rehab and research credibility, and GR-3’s care positioning gives it a clearer institutional wedge than many generic humanoid projects. It is less hot than Unitree or AgiBot because it is not winning on price or viral production scale.
Which robotics startup is most serious about the home today?
1X is the clear answer. Unitree could become a surprise home-adjacent player because of price, but 1X is the only major startup currently forcing the consumer question.
1X opened NEO preorders in late 2025, with first deliveries positioned for 2026, a $20,000 early-access purchase option and a $499-per-month plan. That is the clearest attempt to make a humanoid robot feel like a consumer product. Figure and Apptronik talk more naturally to factories. Agility talks to warehouses. 1X talks to households.
The comparison is brutal for 1X, though. A warehouse robot can move totes in a controlled layout. A home robot has to deal with pets, children, clutter, privacy, fragile objects and unpredictable expectations. That makes 1X more ambitious than proven. Today, it wins the home category because nobody else has made such a direct consumer move, not because the home use case is already solved.
Unitree’s low-cost robots could still matter here. If developers build home-ish applications on cheap hardware, Unitree could enter the consumer conversation from below. But currently, 1X owns the category narrative.
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 humanoid startups are most underpriced by the public conversation?
AgiBot, Neura Robotics and Galbot feel more under-discussed than their evidence suggests. Unitree is no longer under the radar, but many Western readers still underestimate what its price and revenue signals imply.
AgiBot is the clearest under-discussed name. If a U.S. startup claimed a 10,000-unit humanoid milestone in March 2026, it would dominate the market narrative. Because AgiBot is Chinese and less familiar to Western readers, it often gets treated as a secondary name. That feels wrong. Even with caution around the exact number, the company deserves to be in every serious top-startup discussion.
Neura Robotics changed status in June 2026. Before the round, it was easy to treat it as an interesting European robotics company. After up to $1.4 billion in funding, a reported $7 billion valuation and industrial investors across chips, cloud, manufacturing and finance, it became a top emerging humanoid company. We still need deployment proof, but the category changed around it.
Galbot is another name that looks more important once we compare it properly. A RMB 2.5 billion round in March 2026, state-linked capital and reported orders from major industrial groups make it more than a “China also has humanoids” story. It is not as proven as Unitree or AgiBot, but it is one of the strongest emerging signals in China’s embodied-AI layer.
If you want more recent data on this point, please see our latest humanoid robotics market report.
Which famous humanoid startups look most overvalued or under-proven?
Figure AI, 1X, Neura Robotics, Skild AI and Galbot all deserve attention and skepticism at the same time. They are not weak companies; their public proof simply does not yet fully match the scale of their ambition.
Figure is the easiest case. It has the best Western combination of capital and deployment evidence, but a $39 billion valuation prices in an enormous future. Even with BMW and Catalyst Brands, we still do not have public revenue, gross margin or large fleet economics. Figure can be both the current leader and the most demanding valuation story.
1X is under-proven because home robotics is a harder test than enterprise pilots. A preorder is a useful demand signal, but it does not prove retention, safety, autonomy or support economics. The company is exciting precisely because the bar is so high.
Neura’s June 2026 raise is one of the strongest recent signals in humanoids. Still, backlog and production targets should not be confused with delivered robots. Compared with Unitree, Neura has less public manufacturing proof. Compared with most European startups, it is suddenly far ahead.
Skild AI and Galbot have similar evidence gaps in different layers. Skild has an enormous valuation around the robot-brain thesis, but public deployment proof is still thinner than the funding signal. Galbot has impressive financing and order claims, but we need more verified shipment and customer-usage data before ranking it beside Unitree or AgiBot.
This chart, featured in our humanoid robotics market deck, compares the main business model options for humanoid robot manufacturers
Which humanoid startups should we actually put in the top group now?
The current top group is Figure AI, Agility Robotics, Unitree, AgiBot, Apptronik, Neura Robotics, 1X, Skild AI and Physical Intelligence. Galbot is the strongest near-top emerging name, while Fourier Intelligence, Sanctuary AI and Reflex Robotics belong on the watchlist.
Figure is the strongest all-around Western humanoid startup because it appears across capital, deployment and enterprise categories. Agility is the cleanest logistics operator because its tote-moving evidence is more concrete than most humanoid claims. Unitree is the strongest cost-and-commercialization player because it combines low prices, revenue growth and IPO progress. AgiBot is the volume challenger we would watch most closely because its 2026 unit milestone shifts the China conversation from demos to manufacturing scale.
Apptronik is a top U.S. challenger because its capital base and enterprise backers are unusually strong, though it needs more public throughput proof. Neura is now the top emerging European industrial humanoid startup because its June 2026 financing changed its scale overnight. 1X owns the home-humanoid question, even if the home is still the hardest deployment environment. Skild AI and Physical Intelligence should be included because the winning humanoid stack may depend as much on foundation models as on legs and hands.
| Category | Startups selected and why |
|---|---|
| Real-world proof | Figure AI and Agility Robotics. Figure has the strongest public factory-task evidence; Agility has the cleanest logistics throughput metric. |
| Manufacturing scale | Unitree and AgiBot first; Neura Robotics and Apptronik behind. Unitree has revenue and IPO visibility; AgiBot has the loudest unit-scale signal; Neura and Apptronik have capital but less public unit proof. |
| Affordable humanoids | Unitree first, Fourier Intelligence second. Unitree changes the developer-access curve with sub-$20,000 humanoids; Fourier is more institutional at around $125,000. |
| Industrial customers | Figure AI, Agility Robotics, Apptronik and Neura Robotics. Figure has BMW plus Catalyst Brands; Agility has GXO throughput; Apptronik has major enterprise backers; Neura has the strongest European industrial investor mix. |
| Capital momentum | Figure AI, Skild AI, Neura Robotics, Apptronik and Galbot. Figure leads valuation; Skild leads robot-brain funding; Neura is the freshest European capital shock; Apptronik is the U.S. challenger; Galbot is the China emerging capital story. |
| Robot-brain layer | Skild AI, Physical Intelligence, Figure AI and 1X. Skild leads on valuation; Physical Intelligence is the main foundation-model counterweight; Figure and 1X combine hardware with proprietary learning loops. |
| China challengers | Unitree, AgiBot, Galbot and Fourier Intelligence. Unitree leads commercialization; AgiBot leads the unit-volume narrative; Galbot leads the state-backed embodied-AI story; Fourier has the clearest institutional-care wedge. |
| Home humanoids | 1X, with Unitree as a low-cost wildcard. 1X is the only major startup making a direct consumer preorder push today. |
| Under-discussed names | AgiBot, Neura Robotics and Galbot. Each has fresh evidence that is stronger than its current mindshare outside specialist circles. |
| Overvalued or under-proven names | Figure AI, 1X, Neura Robotics, Skild AI and Galbot. All are important, but their public proof still needs to catch up with valuation, backlog, preorder or order-claim narratives. |
| Final top group | Figure AI, Agility Robotics, Unitree, AgiBot, Apptronik, Neura Robotics, 1X, Skild AI and Physical Intelligence. They stand out repeatedly across deployment, scale, affordability, capital, customer access and AI-stack evidence. |
OUR METHODOLOGY
This analysis starts from a simple premise: the answer to “which humanoid startups are really leading?” is not obvious. The market is noisy, fast-moving and often shaped by demo videos, funding headlines and brand familiarity.
To make the answer clearer, we broke the question into separate analytical dimensions rather than treating leadership as one vague, all-purpose ranking.
We looked at where each company showed strength across deployment proof, manufacturing scale, affordability, customer quality, capital momentum, software relevance, China exposure, home robotics and valuation discipline.
For each dimension, we prioritized recent signals that could be compared directly: live operating metrics, revenue visibility, robot pricing, production claims, customer references, funding rounds, valuation data and product availability.
We did not treat every signal equally. Runtime hours, task volume and commercial deployment evidence carried more weight when judging real-world proof.
Revenue, pricing and IPO visibility mattered more for manufacturing and commercialization. Funding and valuation were useful for measuring investor conviction, but not enough on their own to prove operating leadership.
This structured aggregation is what shaped the final group. Instead of asking which startup feels most important, we asked which companies kept appearing across the strongest and freshest evidence categories.
That approach makes the conclusion more grounded, more transparent and less dependent on market noise.
Key sources used for this analysis include: Figure AI on its BMW deployment, Agility Robotics on Digit’s 100,000-tote milestone, Xinhua on Unitree’s revenue and IPO review signal, Unitree’s R1 product page, Unitree’s G1 product page, AgiBot on its 10,000-unit milestone, Figure AI on its Catalyst Brands agreement, GXO on humanoid warehouse strategy, Apptronik on its $935 million-plus Series A, Figure AI on its Series C and $39 billion valuation, Skild AI on its Series C, Business Wire on Skild AI’s funding and valuation, Neura Robotics company materials, The Wall Street Journal on Neura Robotics’ fundraise, 1X’s NEO order page, The Robot Report on 1X’s NEO preorder launch, The Robot Report on Physical Intelligence’s funding, TechNode on Galbot’s RMB 2.5 billion raise, Fourier Intelligence on the GR-3 series, and PR Newswire on Fourier’s GR-3 care-focused humanoid announcement.
This chart, featured in our humanoid robotics market deck, shows the revenue mix across customer segments in the humanoid robotics market
Related blog posts
- Is the Humanoid Robotics Market growing now?
- The most highly valued startups in the humanoid robotics market
- Humanoid robots in factories: for when?
