Japan's Humanoid Revival Is More Significant Than the Hype Suggests
Toyota and Honda are quietly building something the American startups cannot: decades of manufacturing expertise applied to humanoid robots.
By
·25 May 2026·8 min read
The return of Japanese automakers to humanoid robotics deserves more serious analysis than it has received. When Honda announced a new humanoid programme eight years after retiring Asimo, and when Reuters reported that Toyota has accelerated humanoid component procurement, the coverage largely framed this as legacy players catching up to nimble American startups. I think this framing is exactly backwards.
To be precise, what we are witnessing is not Japanese companies belatedly entering a market that Figure and Tesla have defined. We are witnessing the return of organisations with unmatched manufacturing depth to a field they helped create, and their timing may be more deliberate than it appears.
Honda's announcement explicitly references the Asimo lineage, which is interesting because Asimo's reputation has calcified into a kind of shorthand for "impressive demo, no commercial viability." This is not entirely fair. Asimo represented genuine advances in bipedal locomotion, balance recovery, and human-robot interaction. The programme ran from 2000 to 2018, producing multiple generations of increasingly capable systems.
What Asimo lacked was a viable deployment context. Honda built a humanoid that could climb stairs, run, and serve drinks at corporate events, but the company never identified a commercial application that justified the cost. The retirement in 2018 was not a technical failure. It was an admission that the market did not exist.
reports that Honda's new programme emphasises industrial deployment as the commercial target. This is a meaningful shift. The company is not building another Asimo to wave at trade shows. They are building systems intended for factories, warehouses, and (one assumes) automotive assembly lines where Honda has decades of operational knowledge.
Related coverage
More in Humanoids
The headlines are celebrating a $2.5B humanoid robotics deal. I'd pump the brakes a little.
Mark Kowalski · 25 Jun · 6 min
Sometimes the sources don't pan out. Here's what happened when I tried to write a humanoids story this week and ended up with Samsung deals instead.
Sarah Williams · 25 Jun · 3 min
Diffusion models are getting good at imagining robot movements, but 'imaginable' and 'physically possible' aren't the same thing. Researchers are starting to close that gap.
Sarah Williams · 25 Jun · 6 min
A batch of fresh robotics research tackles the same underlying problem from different angles: robots that can see but don't really understand where things are.
I know I'm being picky here, but the distinction matters. A humanoid robot designed from the start for industrial deployment is a fundamentally different engineering challenge than one designed to demonstrate technical capability. The constraints are different. The failure modes are different. The economics are entirely different.
The Reuters report on Toyota is based on supplier interviews, which limits what we can say with confidence. Two suppliers have told the publication that Toyota has accelerated humanoid component procurement in the last six months. The same suppliers serve Honda. One offered a characteristically Japanese framing: "Patience is the strategy."
This is not a lot to go on. We don't know what components, in what volumes, or for what specific applications. Toyota has been working on robotics for years through its Research Institute (TRI), but the company has been notably less public about humanoid development than its American counterparts. The acceleration, if the supplier reports are accurate, suggests Toyota may be further along than its public communications indicate.
Actually, the research shows that Toyota's robotics work has been more substantial than casual observers recognise. TRI has published papers on manipulation, learning from demonstration, and soft robotics. The company acquired Boston Dynamics briefly (before selling to Hyundai) and has maintained robotics partnerships with multiple universities. What Toyota has not done is hold press events with humanoids folding laundry or announce billion-dollar valuations for robotics subsidiaries.
Whether this reflects Japanese corporate culture, genuine strategic patience, or simply a different assessment of the technology's readiness, I cannot say. Probably some combination. But the pattern is notable: both major Japanese automakers are now visibly investing in humanoid robotics after years of relative quiet, and both are framing their efforts in terms of industrial application rather than consumer markets or general-purpose assistance.
The standard narrative positions Tesla, Figure, and other American companies as the leaders in humanoid robotics, with legacy automakers as slow-moving incumbents trying to catch up. This narrative has some basis in reality. The American startups have moved fast, raised enormous sums, and generated significant public attention.
But it's worth noting that the narrative elides some important factors.
First, manufacturing at scale is genuinely hard, and Japanese automakers are genuinely good at it. Toyota's production system is not just a historical achievement. It represents accumulated knowledge about how to build complex electromechanical systems reliably, at volume, with consistent quality. This knowledge does not transfer automatically to humanoid robots, but it transfers more easily than it would to a company that has never manufactured anything physical.
Second, Japanese automakers have extensive supplier relationships and component sourcing capabilities. The Reuters report notes that Toyota's suppliers also serve Honda, suggesting an existing ecosystem of companies capable of producing humanoid components. Building this ecosystem from scratch, as American startups must do, takes time and capital.
Third, and this is speculative, Japanese automakers may have more realistic expectations about deployment timelines. The "patience is the strategy" comment is a single data point, but it aligns with a broader pattern. Honda waited eight years after retiring Asimo before announcing a new programme. Toyota has been working on robotics for years without making aggressive public claims. This could reflect excessive caution, or it could reflect a more accurate assessment of what remains to be solved.
I want to be careful here. I am not claiming that Japanese automakers will necessarily succeed where American startups might fail. The technology challenges are real, the market remains uncertain, and corporate scale can be a liability as easily as an asset. What I am claiming is that the "legacy players catching up" frame significantly underestimates what Toyota and Honda bring to this effort.
Both Honda's announcement and the supplier reports about Toyota emphasise industrial applications. This is the right focus, but it raises questions that neither company has publicly addressed.
The case for humanoid robots in industrial settings rests on a specific claim: that human-shaped robots can operate in environments designed for humans without requiring those environments to be redesigned. A humanoid can (in theory) use the same tools, navigate the same spaces, and perform the same tasks as a human worker, making deployment faster and cheaper than purpose-built automation.
This claim has intuitive appeal, but it hasn't been validated at scale. We don't actually know whether humanoid form factors provide sufficient advantage in real industrial environments to justify their additional complexity and cost. The sample size of deployed industrial humanoids is, to be precise, zero. Everything we think we know about this is based on demonstrations, simulations, and extrapolation.
Honda and Toyota have an advantage here: they own factories. They can test humanoid systems in actual production environments with actual production requirements. They can iterate based on real operational data rather than investor expectations. This is valuable, but it's too early to say whether it will prove decisive.
The methodology question also matters. How do you evaluate whether a humanoid robot is ready for industrial deployment? What metrics matter? What failure rates are acceptable? What tasks should humanoids perform versus purpose-built systems versus human workers? These questions don't have established answers, and the companies that figure them out first will have significant advantages.
Several things remain unclear about Japan's humanoid revival.
First, we don't know the actual state of Toyota's programme. The supplier reports suggest acceleration, but supplier procurement can mean many things. Toyota could be preparing for pilot deployments, or it could be building research prototypes, or it could be stockpiling components for reasons unrelated to near-term deployment. The company hasn't disclosed its timeline or technical approach.
Second, we don't know how Honda's new programme relates to its Asimo heritage technically. The announcement references the lineage, but eight years is a long time in robotics. The field has changed substantially, particularly in AI and machine learning capabilities. Whether Honda is building on Asimo's control systems and mechanical designs or starting fresh with modern approaches would tell us a lot about their strategy.
Third, we don't know how either company is thinking about the AI component. Modern humanoid robotics increasingly depends on learned behaviours, foundation models, and sim-to-real transfer. Japanese companies have historically been stronger in mechanical engineering than in machine learning. Whether they're building AI capabilities internally, partnering with AI companies, or pursuing some hybrid approach matters enormously for their competitive position.
What I'd want to see next: actual technical disclosure. Papers, demonstrations with quantified performance metrics, or at minimum, clear statements about deployment timelines and target applications. The current information environment, a mix of supplier reports and corporate announcements, doesn't allow for serious technical assessment.
I'd also want to see how these companies think about the labour implications. Japan faces severe demographic challenges, with a shrinking workforce and limited immigration. Humanoid robots in factories could address real labour shortages rather than displacing workers. This is a meaningfully different context than the American market, and it might shape both the technology choices and the deployment strategy.
The American humanoid robotics companies are running a particular kind of race: raise capital, generate attention, demonstrate capabilities, raise more capital. This model has produced impressive results and real technical advances. It has also produced incentives to overpromise and to prioritise demonstrations over deployments.
Toyota and Honda appear to be running a different race, one oriented around patient development, manufacturing capability, and industrial deployment. This approach has its own risks. Patience can become paralysis. Manufacturing expertise doesn't guarantee good product design. Industrial deployment is a smaller market than general-purpose humanoids (if general-purpose humanoids ever become viable).
But the return of Japanese automakers to humanoid robotics is more significant than the coverage suggests. These are not companies scrambling to catch up. They are companies with specific capabilities, entering a market at a specific moment, with what appears to be a specific strategy. Whether that strategy succeeds remains genuinely uncertain. But dismissing them as legacy players misses something important about what they actually bring.
The humanoid robotics field is about to get more interesting. And possibly more patient.
