Signals Inbox·July 23, 2026·Humanoid Robotics

Is Humanoid really worth $1.35B today?

Humanoid is not worth $1.35 billion on current financials, but its factory trials, Schaeffler contract, Bosch manufacturing support and 34,000-robot pipeline make the valuation defensible as a very aggressive venture bet.

Q1What happened in Humanoid's $1.35 billion funding round?

Humanoid's new $1.35 billion valuation is a real price paid by outside investors before the company has begun its first commercial beta rollout.

In its latest announcement, the London robotics startup said it had raised $152 million in a Series A led by Prime Movers Lab, with Schaeffler, Bosch, Fubon Financial Holding Venture Capital and Aglaé Ventures also participating. The company says the round takes total funding to $270 million. Because the $1.35 billion figure is post-money, the implied pre-money valuation is about $1.20 billion, and the new investors collectively bought roughly 11% of the business.

The speed is unusual. Artem Sokolov founded Humanoid in 2024 and told Reuters earlier this year that he had personally invested about $100 million before bringing in external shareholders. Within roughly two years, Humanoid hired more than 250 people, opened offices in London, Boston, Vancouver and San Diego, ran factory trials with large industrial groups and became a unicorn. Agility Robotics needed more than a decade to reach its proposed $2.5 billion public-market transaction, while Figure reached a far higher valuation on an equally aggressive timeline.

That pace helps explain the funding, although it also shows how much future performance has already entered the price. Humanoid plans to use the cash for beta deployments, mass manufacturing of its wheeled platform, a new robot generation and further development of KinetIQ, its robotics AI system. None of those programmes has yet produced a disclosed stream of commercial revenue.

Humanoid's funding and valuation structure

Q2Does Humanoid have real revenue today?

Humanoid currently has no disclosed revenue, annual recurring revenue or commercial run rate that can independently support a $1.35 billion valuation.

The company talks publicly about proofs of concept, future deployments and pre-orders, but it has not published recognized revenue, paid robot counts, customer retention or gross margin. Its own financing announcement says beta robots will begin rolling out in the fourth quarter of 2026, followed by mass manufacturing of wheeled units. That wording places Humanoid at the start of commercialization rather than inside a mature revenue ramp.

Some money may already be coming in through engineering work, deposits or paid trials, but Humanoid has not disclosed those figures. Factory tests with Siemens, Ford, Bosch or Schaeffler also cannot be treated automatically as recurring revenue. A customer can fund a pilot, share engineering costs or provide a test site without committing to a large paid fleet.

For now, Humanoid belongs in the pre-revenue or very early-revenue category. Its valuation rests mostly on contract conversion, manufacturing capacity and technical progress that investors expect to see next.

Q3Are Humanoid's 34,000 robot pre-orders worth $2.4 billion?

Humanoid's 34,000 pre-orders show serious customer interest, although the claimed $2.4 billion remains a projection until robots are delivered and customers start paying.

Artem Sokolov told Reuters that deliveries were planned over three years and could represent roughly $2.4 billion in future annual recurring revenue. The arithmetic implies about $70,600 of annual revenue per active robot. That price could make sense for a Robot-as-a-Service package covering the machine, software, maintenance, updates and support. That is not nothing.

The missing contract details change everything. Humanoid has not disclosed how many reservations carry non-refundable deposits, what customers can cancel, whether the orders contain deployment conditions, or how concentrated the pipeline is. Industrial buyers often test several suppliers before choosing one platform, so the same factory expansion may appear in more than one vendor's pipeline.

The number works best as a ceiling on possible demand. If every reservation converts, Humanoid has more potential business than its present valuation suggests. If only 5% converts, the fleet would still contain 1,700 robots and imply about $120 million of annual revenue at the company's stated economics. The valuation can work with a modest fraction of the headline pipeline, but we still need proof that those customers will sign final contracts and accept delivered machines.

Q4What revenue multiple are Humanoid's investors paying now?

Humanoid's current revenue multiple is unknowable, and any precise figure would create a false sense of accuracy.

If recognized revenue is close to zero, the price-to-sales ratio is effectively enormous. If we divide the valuation by the company's $2.4 billion future-ARR claim, the result is only 0.56 times revenue, which would make Humanoid look absurdly cheap. Both calculations mislead because one ignores expected growth and the other treats unconverted reservations as active subscriptions.

We get a more useful answer by working backwards from reasonable forward multiples. At 30 times revenue, Humanoid needs $45 million of annual sales. At 20 times, it needs $67.5 million. At 15 times, the requirement rises to $90 million. These are achievable levels for a successful industrial robot provider, although Humanoid has not reached them publicly.

In plain English, investors expect Humanoid to turn several hundred robots into recurring contracts quickly and then scale into the thousands. The round looks extreme against present revenue, but it becomes easier to defend against a credible 2028 or 2029 run rate.

Q5How much of Humanoid's demand is backed by a real contract?

Humanoid's Schaeffler agreement is real, but one customer contract cannot carry a $1.35 billion valuation by itself.

Humanoid and Schaeffler describe their arrangement as a binding, phased deployment and supply agreement. The first systems are due to enter live German production before the end of 2026, with work across two sites continuing into 2027. Schaeffler ultimately targets a four-digit number of wheeled robots across its factories by 2032.

The agreement goes much further than a standard memorandum of understanding. Humanoid will provide robots through a recurring service model that includes fleet software, maintenance, updates, technical support and performance management. Schaeffler also agreed to become the preferred actuator supplier for more than half of Humanoid's wheeled-platform demand through 2031, with the companies expecting a seven-digit actuator volume over five years.

The published material gives no minimum revenue, deployment price, cancellation penalties or guaranteed robot count. A four-digit target can mean 1,000 units or nearly 10,000, and the first sites still have to prove that the robots work reliably at an acceptable cost. Schaeffler gives Humanoid real proof that a major industrial buyer wants the product, though one buyer only answers part of the valuation question.

Q6Are Humanoid's factory trials good enough for paying customers?

Humanoid's factory trials are good enough to keep customers moving forward, but they have not yet shown fully unattended industrial work.

At Ford's innovation centre in Cologne, Humanoid reported 97% autonomous pick-and-place reliability and 83 picks per hour, comfortably above the trial target of 50. The robot also ran continuously for one hour, twice the original target. Those results came from handling totes and metal parts in a real automotive setting, which gives them more weight than a controlled stage demonstration.

Siemens later reported that HMND 01 Alpha completed more than eight hours of autonomous tote handling at its Erlangen electronics factory. The robot moved 60 totes per hour and exceeded 90% pick-and-place success. An eight-hour shift is useful evidence of endurance, although a failure rate approaching one action in ten would still create too many interruptions for many production lines.

Humanoid's more recent KinetIQ work produced better internal results. On a bimanual tote task, the company says success increased from 77.6% to 98.9% while throughput rose from 122 to 279 cycles per hour. That is rapid progress, yet 98.9% still means roughly 11 failed cycles per 1,000 attempts. A robot working thousands of cycles per day needs reliable recovery from those failures, otherwise a human remains close by.

The trials already show useful work at relevant speeds. Commercial readiness now depends on repeatability across weeks, different sites and changing objects, plus clear data on human interventions and maintenance.

Q7Can Humanoid manufacture 34,000 robots in three years?

Humanoid has not yet shown that it can deliver 34,000 robots in three years, although Bosch and Schaeffler give the company a better chance than most young robotics startups.

Fulfilling that pipeline would require an average of about 11,300 robots per year, 944 per month or 31 every day. Agility Robotics spent years building RoboFab, a dedicated facility designed for up to 10,000 Digit robots annually. Unitree shipped about 5,500 humanoids in 2025 after developing its manufacturing base across several robot categories. Humanoid would need to pass both scales very quickly.

Bosch has agreed to manufacture HMND 01 robots for Europe and help with hardware design, supply chains, serviceability, reliability and cost. Schaeffler covers a large part of the actuator requirement. These partnerships reduce the usual risk of a startup trying to invent a robot and a factory at the same time.

Those partnerships still leave Humanoid responsible for stable designs, certified safety systems, final testing, installation teams, spare parts, field support and customer acceptance. XPeng's current plan to reach more than 1,000 humanoid robots per month by the end of 2026 shows that this production rate is entering the industry conversation, but XPeng can reuse a large automotive supply chain and manufacturing workforce.

Humanoid's reservation total creates a useful demand target. We would only treat the full delivery schedule as credible after the company publishes monthly output, accepted deployments and factory capacity.

Q8Does Humanoid's wheeled robot have an advantage over bipedal rivals?

For factories today, Humanoid's wheeled platform is probably the right commercial choice because most customers gain little from paying for bipedal walking.

Most warehouses and production plants already use flat floors, marked routes, ramps and standardized workstations. Wheels consume less energy, remove the balance challenge and lower the chance of a costly fall. Humanoid can devote more battery capacity and engineering effort to arms, hands, perception and task execution, where the customer receives the economic benefit.

The Ford, Siemens, Bosch and Schaeffler work all centres on practical jobs such as tote movement, box handling, parts transfer and machine feeding. Walking would add spectacle without necessarily improving those workflows. This focus could let Humanoid generate revenue while bipedal competitors continue refining locomotion.

HMND 01 also faces a wider competitive set. It must beat autonomous mobile robots fitted with arms, fixed automation cells, collaborative robots and humanoids from Figure, Apptronik or Agility. A factory may decide that a less human-looking machine can complete the same job more cheaply and reliably.

Humanoid's design choice shortens the route to usefulness. Its pricing power will come from flexible manipulation and fast deployment across many tasks, not from the robot's body shape.

Q9Can competitors easily copy Humanoid's KinetIQ software?

KinetIQ could become Humanoid's main moat, although the available evidence currently shows fast technical improvement more clearly than a defensible software network.

Humanoid describes KinetIQ as a four-layer framework covering fleet orchestration, task planning, learned manipulation and low-level control. Its KinetIQ Ascend programme trains robots through real-world reinforcement learning after imitation learning provides the starting behaviour. The company reported a 42% throughput increase when feeding bearing rings, an 85% increase on an object-handover task and the sharp tote-handling improvement described earlier.

Traditional factory automation often requires engineers to program each motion and redesign the system whenever the task changes. A robot that improves through practice and transfers skills between sites would lower deployment costs and create a growing data advantage.

Humanoid has not yet disclosed the size of its training dataset, the amount of human teleoperation required, or how well a skill learned in one factory transfers to another. Figure, Apptronik, NEURA, Tesla and several Chinese developers are also building large embodied-AI systems with access to substantial capital and data.

KinetIQ becomes a real moat when new customer deployments get faster, fleet reliability improves with every installed unit and switching platforms becomes painful for customers. Company-reported task benchmarks have not proved that advantage yet.

Q10Is Agility Robotics a better valuation benchmark for Humanoid?

Agility Robotics makes Humanoid's valuation look demanding because Agility is worth less than twice as much while showing far more operating history.

Agility's proposed public transaction values the company at $2.5 billion before new capital. Its investor materials report more than 65,000 operating hours across nine facilities, over $300 million of conditional multi-year orders and named deployments with Schaeffler, GXO, Toyota and Mercado Libre. RoboFab is designed for annual capacity of up to 10,000 Digit robots.

Humanoid is priced at 54% of Agility's transaction value. That ratio feels high for a company with no comparable fleet operating history, especially when Agility has commercial hours and manufacturing infrastructure that investors can inspect. Humanoid's larger claimed pipeline partly closes the gap, but reservations carry less weight than conditional orders tied to an established product.

The discount could narrow over time. Humanoid is moving with a lighter wheeled design, has Bosch handling contract manufacturing and can focus on recurring industrial service contracts from the start. If that model produces faster installations and lower hardware costs, the company may compress several years of Agility's learning curve.

Today, Agility is the fairer benchmark. Humanoid's valuation assumes that faster execution will compensate for a large maturity gap.

Q11Does Unitree make Humanoid look overpriced?

Unitree makes Humanoid look expensive on current fundamentals because Unitree already combines real shipments, revenue and profit.

The Financial Times reported from Unitree's IPO materials that the Chinese company generated about RMB1.7 billion, roughly $250 million, of revenue in 2025. Unitree shipped approximately 5,500 humanoid robots, produced an adjusted profit near RMB600 million and recorded a gross margin close to 60%. A possible RMB40 billion valuation would equal about $5.9 billion, or roughly 23 times 2025 revenue.

Humanoid is currently worth about 23% as much as Unitree. At Unitree's revenue multiple, Humanoid would need close to $59 million of annual sales, equivalent to roughly 830 active robots at the implied $70,600 annual service price.

Unitree's revenue quality needs context. Its prospectus indicates that much of humanoid demand still comes from universities, research institutions and display uses, with genuine industrial applications forming a small share. Humanoid is aiming at harder factory tasks and potentially more valuable recurring contracts.

Even with that distinction, Unitree sets a high bar. Humanoid needs unusually strong customer and production evidence before receiving almost one-quarter of the value assigned to a profitable manufacturer with thousands of shipments.

Q12Do Figure, Apptronik and NEURA make Humanoid look cheap?

Figure, Apptronik and NEURA make Humanoid look cheap only inside a private humanoid market where investors are already pricing enormous future outcomes. Cheap is doing a lot of work in that sentence.

Figure announced more than $1 billion of Series C commitments at a $39 billion post-money valuation. Apptronik has raised more than $935 million in its Series A and extension, with reporting placing the company above $5 billion. NEURA announced a Series C of up to $1.4 billion, while the Financial Times reported a valuation near $7 billion and the company disclosed an orderbook and deployment pipeline above $1 billion.

Humanoid sits far below those prices. Figure is valued almost 29 times higher, NEURA roughly five times higher and Apptronik about four times higher. The gap makes sense because the peers have broader investor bases, longer development histories or deeper strategic programmes. Apptronik works with Google DeepMind, Mercedes-Benz, Jabil and John Deere. NEURA already sells several robot types and is building a shared learning ecosystem. Figure is spending heavily on its Helix AI system, manufacturing and data collection.

These rounds show that investors will fund scarce humanoid teams long before revenue matures. They offer little protection if enthusiasm cools. Once investors start focusing on delivered fleets and gross profit, several private valuations could fall together.

Humanoid compared with leading humanoid robotics companies

Q13What do public robotics companies say about Humanoid's valuation?

Public robotics companies suggest Humanoid needs roughly $90 million to $120 million of credible forward revenue before $1.35 billion starts to look comfortable.

Symbotic currently has a market value near $5.8 billion. Its latest reported quarter produced $676 million of revenue, up 23%, which gives an annualized run rate of about $2.7 billion and a market-cap multiple close to 2.1 times. Symbotic is a mature warehouse-automation provider, so Humanoid deserves a higher growth multiple if deployments take off.

Serve Robotics offers the speculative end of the range. Serve has a market value around $405 million and guides to approximately $26 million of revenue for 2026, equal to roughly 15.6 times forward sales. Serve also expects very high operating expenses, which shows how generously public investors can value an early robot fleet when growth expectations are strong.

Intuitive Surgical provides a profitable, high-quality robotics benchmark. Its market value is around $125 billion and trailing revenue is roughly $11 billion, a multiple near 11 times. Intuitive earns recurring revenue from instruments and service on a large installed base, exactly the kind of revenue quality Humanoid hopes to build eventually.

Applying the Serve multiple to Humanoid implies about $87 million of forward revenue. Applying Intuitive's multiple gives roughly $119 million. Humanoid can justify a higher figure during hypergrowth, but current public markets still expect measurable sales, installed units and margin evidence.

Q14Is the humanoid robotics market big enough for Humanoid now?

The humanoid robotics market is large enough for Humanoid to become a billion-dollar company, although today's paid industrial demand remains far below the industry's biggest forecasts.

The underlying automation market is already large. The International Federation of Robotics recorded 542,000 industrial robot installations in 2024, more than twice the annual number a decade earlier. Factories clearly spend on automation when the return is visible, and humanoids could unlock tasks that fixed arms or conventional mobile robots struggle to handle.

Current humanoid activity is accelerating quickly. Chinese manufacturers produced an estimated 85% of global humanoids in 2025, and industry forecasts expect Chinese shipments to reach roughly 50,000 units in 2026. XPeng now aims to exceed 1,000 units per month by the end of the year. At the same time, reporting on Unitree's sales mix suggests that only a minority of present humanoid demand comes from genuine industrial work.

The latest labour reaction also shows that customers and workers are taking the technology seriously. Hyundai's Ulsan plant recently faced a partial strike linked to concerns about future Atlas deployments, while Hyundai plans to introduce the Boston Dynamics robot at its Georgia factory by 2028. That dispute says little about near-term robot revenue, but it confirms that large manufacturers are planning beyond small laboratory trials.

Forecasts remain extremely wide, from tens of billions of dollars in the 2030s to multi-trillion-dollar long-term scenarios. Humanoid needs neither extreme. A few thousand well-used factory robots could support its current price. The market risk lies in adoption speed: industrial customers move slowly, demand high uptime and often find simpler automation alternatives.

Q15What revenue would justify Humanoid's $1.35 billion valuation?

Humanoid needs at least $70 million to $100 million of forward annual revenue for the valuation to look plausible without relying on the sector's most speculative multiples.

At the company's implied service revenue of about $70,600 per active robot, a 20-times multiple requires roughly 956 deployed units. A 15-times multiple needs about 1,275 units, while a more disciplined 10-times multiple needs approximately 1,913. These levels are small compared with the claimed pipeline and still large compared with Humanoid's disclosed commercial fleet today.

A 30-times multiple could apply briefly if revenue triples, retention is strong and hardware margins improve. Once growth slows, the market will probably move closer to the public robotics range. The practical target is therefore more than a single revenue milestone: Humanoid must reach it with repeatable installations, attractive service economics and limited human support.

Revenue needed to support a $1.35 billion valuation

Q16What would make the bull case for Humanoid work?

Humanoid can grow into $1.35 billion if its industrial partnerships turn into roughly 1,000 paid robots quickly and KinetIQ lowers the cost of every new deployment.

The company has assembled an unusually useful group of partners for its age. Siemens and Ford have tested real tasks. Schaeffler has agreed to a phased commercial rollout and supplies actuators. Bosch supports manufacturing. NVIDIA provides part of the computing and simulation stack. The new funding gives Humanoid enough capital to build inventory, hire deployment teams and absorb early hardware losses.

The wheeled-first strategy also helps. Humanoid can sell tote handling, parts movement and machine feeding without waiting for perfect walking. If customers can move the same robot between several tasks, the service fee could replace multiple pieces of specialized equipment and justify a healthy recurring price.

The bull case becomes convincing when three things appear together: several hundred accepted robots, task reliability near the company's 99.9% target and evidence that later deployments take less time than the first ones. Reaching $70 million to $100 million of annual revenue with those characteristics would make the valuation reasonable and leave room for further growth.

Q17What could destroy Humanoid's valuation?

Humanoid's valuation would unravel quickly if customer reservations convert poorly, factory reliability stalls or service costs make the recurring revenue model look like expensive equipment leasing.

Contract conversion is the first risk. Customers may reserve several competing robots, delay factory spending or cancel when final economics become clear. A pipeline that appears huge before delivery can shrink sharply once buyers must pay deposits, change workflows and accept service commitments.

Reliability comes next. The best disclosed internal task reached 98.9% success, while the Siemens trial exceeded 90%. Those figures are promising for development and still leave many daily failures at production volume. Slow recovery, frequent human intervention or safety stops would weaken the labour-saving calculation.

The business model also carries more capital intensity than software. Humanoid must manufacture machines, finance working capital, maintain fleets, replace parts and provide round-the-clock support. If gross margins stay low, investors will stop applying software-like revenue multiples.

Competition could squeeze the remaining upside. Figure, Apptronik, Agility, NEURA, Tesla, Boston Dynamics and Chinese manufacturers are all chasing overlapping factory tasks. Conventional automation firms can also combine mobile bases, robot arms and vision systems. Humanoid must become clearly cheaper, faster to deploy or more flexible than those alternatives.

Q18Is Humanoid really worth $1.35 billion today?

At $1.35 billion, Humanoid looks aggressively priced but still defensible as a venture bet. The company's disclosed business has not yet earned that value on normal financial measures.

The evidence supporting the price goes beyond the usual humanoid demo reel. Humanoid has completed relevant factory tests, secured a binding industrial rollout, lined up an experienced manufacturing partner and built a team of more than 250 people in two years. Its wheeled platform also targets jobs that customers can automate now, without waiting for human-level locomotion.

The problem is easy to state. Humanoid discloses no current revenue, has not begun its commercial beta rollout and still reports task reliability below the standard needed for unattended work. The 34,000 reservations could make the valuation look cheap if they convert, but the company has published too little information about deposits, cancellations and customer concentration.

Our answer is a cautious yes as a venture bet and a clear no on current financials. Humanoid needs roughly 1,000 paid robots, around $70 million to $100 million of forward revenue, strong retention and improving margins within the next few years. It also needs to show that KinetIQ makes each deployment faster and more reliable than the last.

For now, investors own a well-funded route to a valuable industrial robotics business. Humanoid has enough real evidence to avoid the “pure hype” label, but the latest round has priced in several milestones before the company has reached them.