Expert Insights on Humanoid Robots
On July 9, 2026, Mitsubishi Motors of Japan officially announced its entry into the humanoid robot track, reaching a deep mass production cooperation with Highlanders, a local robotics startup incubated by the University of Tokyo, officially joining the ranks of automotive giants crossing over into the field of embodied intelligence.
In-House Developed Humanoid Robots: Dual Strategy for Internal Use and External Sales
According to officially disclosed plans, the humanoid robot jointly developed by both parties will achieve mass production at Mitsubishi's Kyoto engine plant in 2027, with a planned monthly capacity of 1,000 units. The products will be prioritized for Mitsubishi's own vehicle and engine plants, taking over repetitive industrial operations such as parts transportation, engine assembly, and material sorting. After on-site factory testing and meeting performance and stability standards, they will be open for external sales to manufacturing enterprises across the industry.
As early as May 2026, Mitsubishi Motors had completed a strategic investment in Highlanders. Takahiro Kato, President of Mitsubishi, stated that in the future, additional investments will be made at appropriate times based on R&D progress and mass production implementation, steadily deepening the technological binding and industrial synergy between the two parties.
In this cooperation, Mitsubishi will export its mature vehicle manufacturing processes, comprehensive automotive supply chain system, large-scale mass production control experience, and industrial scenario implementation capabilities; Highlanders will provide core technologies in humanoid robot hardware architecture and physical AI large models. Both parties will carry out exclusive iterative optimization for high-frequency operation scenarios in automotive manufacturing, creating dedicated humanoid robots adapted to high-intensity industrial working conditions.
Multiple Automakers Cluster for Mass Production in 2027
Currently, the humanoid robot track for global automakers has entered the sprint stage of mass production. The industry shows a clear characteristic of concentrated mass production clustering in 2027. The commercialization landing nodes of leading players highly overlap, ushering in a window period for concentrated capacity release.
Tesla's Optimus plans to complete internal small-batch deployment in 2026 and fully launch external commercial deliveries in 2027. Relying on the advantage of reusing autonomous driving FSD technology, it will cater to both industrial production and general living scenarios.
XPeng IRON and GAC GoMate Mini both adopt the rhythm of "trial production and iteration in 2026, scale mass production and external sales in 2027," focusing on segmented tracks such as automotive production lines, warehousing and logistics, and commercial services.
BMW, in partnership with a U.S. startup, will launch pilot applications in factories in 2026 and continuously expand the scope of production line deployment in 2027.
Hyundai Motor's pace is relatively conservative, delaying its main large-scale mass production to 2028, actively staggering the industry's clustering cycle.
Overall, 2027 will become a critical year for global automakers' humanoid robots to transition from prototype demonstrations to large-scale commercial use.
Mass Production Wave to Trigger Multiple Industry Reshuffle Risks
The collective clustering of mass production by automakers appears to be a concentrated explosion of industry technological maturity, but it actually hides structural risks such as overcapacity, homogenized competition, and insufficient scenarios. This may completely rewrite the industry competition landscape and initiate a large-scale industry reshuffle.
Currently, most humanoid robots mass-produced by automakers can only complete standardized fixed processes such as transportation, sorting, and simple assembly, and cannot adapt to complex, unstructured industrial operating environments. They suffer from pain points such as weak intelligent versatility and expensive operation and maintenance costs. After concentrated mass production in 2027, hardware capacity will climb rapidly, but the growth rate of mature commercial scenarios that can be landed and high-quality customer orders will be far behind the capacity expansion speed, inevitably causing product inventory backlog. To compete for limited market share, enterprises will be forced to engage in low-price competition, leading to a significant decline in the overall gross profit margin of the industry, and being forced to rely on vehicle business for financial support and subsidies. At that time, followers lacking core technologies and relying solely on assembly and OEM will be the first to be eliminated by the market.
Influenced by hot industry expectations, upstream enterprises for general components such as reducers, servo motors, sensors, and embodied AI chips have expanded production and stocked up in advance. After centralized delivery in 2027, they may face the dilemma of oversupply, plummeting prices, and shrinking profits, thereby compressing investment in cutting-edge technology R&D. However, high-end core components such as high-reliability joints, long-life dexterous hands, and high-precision servo systems adapted to high-intensity industrial operations have long-term capacity gaps due to high technical barriers and difficult R&D. This has become a core bottleneck restricting the performance upgrade of the whole machine, seriously dragging the high-end development process of the industry.
After the reshuffle stage of clustered mass production, the pace of survival of the fittest in the industry will be significantly accelerated. Leading automakers equipped with full-stack embodied intelligence algorithms, closed-loop data from real factory scenarios, and automotive-grade mass production quality control capabilities will break away from the traditional model of single hardware sales. Relying on stable product performance and deep scenario adaptation capabilities, they will derive multiple profit models such as robot leasing, on-demand operation and maintenance, AI model subscription, and overall solutions for production line automation, building deep technological and service barriers. In contrast, participating enterprises lacking self-developed algorithms, scenario iteration capabilities, and relying solely on hardware assembly will find it difficult to form core competitiveness due to severe product homogenization and insufficient stability, and will eventually gradually exit the track.
After a large number of humanoid robots are batch-introduced into manufacturing production lines, problems such as human-robot collaboration collisions, program failures, positioning deviations, and data security will break out in a concentrated manner, which is highly likely to trigger operational risks such as production line shutdowns and production accidents. At the same time, the communication protocols, safety standards, and adaptation specifications of products from various brands are not unified, resulting in high costs and poor compatibility for cross-factory reuse of equipment. This will prompt countries to accelerate the introduction of safety operation specifications, human-robot collaboration standards, and AI data compliance systems for humanoid robots, significantly raising industry entry barriers and thoroughly rectifying the chaos of unchecked growth in the industry.
Previously, the humanoid robot industry has long relied on valuation based on future narratives, with obvious bubbles. After mass production lands in 2027, the reality of corporate profits falling short of expectations and continuous losses will end the speculative hype logic. The focus of market investment will shift from the "capacity concept" to fundamentals such as core technologies, scenario landing, and sustainable profitability. The heat of industry investment and financing will cool down periodically, the survival pressure on startups will increase, the M&A and integration of the industrial chain will accelerate, and the industry dominance of leading automotive giants will be further consolidated.
Broad Long-Term Space in the Track, Industry to Enter High-Quality Development
From the perspective of long-term industry fundamentals, the growth potential of the humanoid robot track remains broad. Data from Germany's Roland Berger shows that in 2025, the market size of humanoid robots for OEMs was less than $1 billion. Optimistically, it is expected that by 2035, the market size will surge to $750 billion, ushering in exponential growth over a decade.
The global industrial competition landscape is becoming increasingly fierce. In 2025, 85% of the globally shipped humanoid robots were produced by Chinese manufacturers. China has clearly proposed the goal of achieving global leadership in the comprehensive competitiveness of the robot industry by 2035; Japan has also planned to deploy 10 million robots by 2040, continuing to support local core technology R&D. The dual dividends of policy and market will escort the long-term development of the industry.
The clustered mass production in 2027 is a critical testing ground for humanoid robots to move from concept speculation to commercial landing. Mitsubishi Motors' steady approach of "internal verification first, external sales later" can rely on its own factories to continuously polish products, accumulate scenario data, and optimize iterative technologies, minimizing commercial trial-and-error costs and effectively hedging against the risk of industry homogenized competition. In the future, whether they can break through the three core pain points of general embodied AI, long-life hardware, and low-cost mass production, and whether they can achieve large-scale scenario landing and sustainable profitability, will become the core confidence for automakers to stand firm in the humanoid robot track.
After the reshuffle in 2027, the humanoid robot industry will completely bid farewell to extensive expansion and officially enter a new stage of high-quality development driven by technology and prioritizing value.
Data sourced from publicly available information on the internet. Corrections and supplements are welcome. This article represents only the author's views and does not constitute any investment advice.