Manufacturing
Mitsubishi and Highlanders Put Humanoid Production on the Old Engine Plant Clock
Mitsubishi Motors' Highlanders deal turns an unused Kyoto engine facility into a test case for whether automakers can become humanoid robot manufacturers.
Mitsubishi Motors and University of Tokyo spinout Highlanders are exploring humanoid robot production at Mitsubishi's Kyoto Plant, with outside reports pointing to a target of up to 1,000 units per month if the plan reaches scale.
The companies announced the memorandum of understanding on July 9, 2026, but the story gained fresh industry attention this week because it fits a pattern that keeps getting harder to ignore: automakers are becoming the manufacturing layer for physical AI. Mitsubishi is not just talking about using robots. It is studying whether an unused part of an engine plant can become a humanoid production base as early as 2027.
AI-generated image
The Kyoto Plant plan matters because it connects idle automotive capacity with a new robotics manufacturing thesis. Source: AI-generated editorial illustration.
Key Stats
2027
Earliest Production Feasibility
1,000/mo
Reported Scale Target
May 2023
Highlanders Founded
Kyoto
Plant Under Review
The News
Mitsubishi Motors says it has signed an MOU with Highlanders, Inc. to collaborate on a new industrial foundation where humans and robots work together. The agreement has two linked tracks. First, the companies will explore joint development of humanoid robots for Mitsubishi's own manufacturing facilities. Second, Mitsubishi will study mass production of Highlanders products at its Kyoto Plant.
Mitsubishi's official announcement says the company will examine whether production can begin in early 2027 by using currently unused buildings at the Kyoto site. It also says Mitsubishi has already invested in Highlanders and plans to make more investments. That matters because this is not a loose marketing tie-up. It is an automaker putting capital, plant space, quality assurance knowledge, mechatronics control experience, and factory operations into a humanoid robotics startup's scale-up problem.
Electrek and industry aggregators amplified the story this week with the most aggressive interpretation: Mitsubishi could build up to 1,000 Highlanders humanoid robots per month at a former internal combustion engine facility. That number should be treated as a target, not a proven capacity. Even so, it is the kind of number that changes how investors, suppliers, and industrial buyers read the humanoid market.
The companies have not published a full spec sheet for a finished production robot, a customer rollout schedule, or a confirmed purchase order. The confirmed part is narrower but still important. Mitsubishi wants to test humanoid robots in its own manufacturing operations, learn from that practical use, and evaluate whether it can become a production partner for Highlanders.
Why It Matters
Humanoid robotics is shifting from lab credibility to manufacturing credibility. Mitsubishi gives Highlanders access to a plant-floor partner that knows durability, safety design, quality assurance, and repeatable production, the exact disciplines that determine whether physical AI becomes a shipped product or remains a prototype category.
The Old Engine Plant Signal
The strongest part of the Mitsubishi story is not the phrase humanoid robot. It is the phrase unused buildings at the Kyoto Plant. Across the automotive sector, the move from internal combustion systems to electrified platforms is leaving some factories with awkward assets: trained workers, machine tools, logistics routines, safety systems, and floor space that may no longer map cleanly to future vehicle demand.
Humanoid robotics gives automakers a possible second life for that capacity. A humanoid robot is not a car, but the production challenge has familiar pieces. It needs electromechanical assembly, supplier control, test procedures, calibration, endurance validation, repair loops, and strict defect tracking. It also needs cost discipline. Those are not natural strengths for a pure AI lab. They are the everyday business of an automaker.
Mitsubishi's public language is careful. The company says it will explore production and practical use, not that it has already opened a robot mega-factory. That caution is healthy. The humanoid industry has been full of production promises that arrive before the robot can work eight hours in a messy facility. Still, the Kyoto angle is concrete enough to deserve attention because it attaches humanoid ambition to physical manufacturing infrastructure.
Idle Capacity
Unused plant buildings can become strategic assets if robot production uses automotive-grade process control.
Internal Test Site
Mitsubishi can deploy robots in its own factories before treating outside customers as the proving ground.
Supplier Gravity
A credible 2027 production path would pull actuator, sensor, battery, and controls suppliers toward Highlanders.
AI-generated image
The hard part of humanoid scale is repeatable manufacturing, inspection, and calibration, not the launch photo. Source: AI-generated editorial illustration.
What Highlanders Gets From Mitsubishi
Highlanders was established in May 2023 and describes its business as development of general-purpose humanoid and quadruped robots, robot simulators, and related services. Drives & Controls reported that the startup is developing a physical intelligence foundation model called Kepler and a fleet of robots that can generate real-world training data for that model.
For a young robotics company, that software ambition needs a hardware flywheel. Robots have to be built, deployed, serviced, upgraded, and returned to work after failures. Data quality depends on where the robots operate and how consistently their sensors, actuators, and control systems behave. A messy build process produces messy learning signals. A disciplined production partner can tighten that loop.
Mitsubishi brings several pieces Highlanders would struggle to build quickly alone. The automaker cites mass-production engineering, quality assurance, durability and safety design, integrated mechatronics control technologies, and factory operations. Each one maps to a known humanoid bottleneck. Actuators need repeatable performance. Battery packs need safe service procedures. Perception modules need calibration. Whole-body control software needs predictable mechanical behavior. Field support needs documentation and spares.
That does not make Highlanders a winner by default. It does make the startup more interesting than a company trying to leap from university lab to global factory supplier with no industrial partner. In humanoids, execution credibility is starting to matter as much as demo quality.
| Question | Confirmed By The MOU | Still Unproven | Why It Matters |
|---|---|---|---|
| Production Site | Kyoto Plant buildings are under review for production feasibility. | Final line layout, capex, and launch approval. | A real plant gives the plan more weight than a prototype lab. |
| Deployment | Mitsubishi plans practical use in its own manufacturing facilities. | Robot count, task list, safety case, and uptime data. | Internal use can reveal whether the robots create operational value. |
| Scale | The companies are studying mass production of Highlanders products. | The reported 1,000-unit monthly target has not been proven in output. | Scale claims need delivered units, not planning language. |
| AI Stack | Highlanders frames its work around robotics and physical AI. | Model performance in factory tasks remains undisclosed. | Physical AI value depends on reliable task completion under constraints. |
Why Automakers Keep Showing Up
Mitsubishi is not alone in seeing humanoids through an automotive lens. Hyundai owns Boston Dynamics and is preparing Atlas for industrial work. BMW has run Figure pilots. Toyota appears in Agility's customer orbit. Chinese automakers and suppliers are also close to the category, either as customers, investors, or manufacturing partners. The reason is simple: car plants contain many dull, repetitive, ergonomically awkward jobs that are still hard to automate with fixed machines.
Humanoids promise a different route. Instead of redesigning every workstation around a custom automation cell, a factory may eventually place a mobile, sensor-rich, two-armed machine into spaces built for people. That is the theory. The practical version is narrower. Early robots will likely handle structured logistics, machine tending, inspection support, kitting, and simple material movement before they take on dense assembly work.
That is why Mitsubishi's internal deployment plan matters. If the company puts Highlanders robots into its own facilities, the first question should not be whether they look human. It should be what task they perform, how long they run, how many interventions they need, how they are supervised, and whether the factory wants more after the pilot.
Deployment Reality Check
AI-generated image
Automotive plants are becoming the proving ground for whether general-purpose robots can fit existing industrial workflows. Source: AI-generated editorial illustration.
The 12-Month Outlook
The next year should answer three questions. First, does Mitsubishi move from feasibility language to a real line plan at Kyoto? Second, does Highlanders disclose enough hardware and software detail to be compared with other industrial humanoid platforms? Third, do the companies show internal factory tasks with measurable performance, not just collaboration photos?
For now, the honest read is mixed. The Mitsubishi partnership is a strong signal because it brings manufacturing depth to a young robotics company. The lack of disclosed robot counts, factory tasks, purchase terms, and performance metrics keeps it from being proof of commercial adoption. In other words, this is credible infrastructure, not completed deployment evidence.
That distinction matters for the wider market. Humanoid robotics is entering the phase where announcements must be sorted by the quality of the operating evidence behind them. A lab demo says a robot can move. A pilot says it can enter a worksite. A production plan says someone thinks it can be manufactured. A paid, repeatable deployment says the economics are starting to work. Mitsubishi and Highlanders have moved into the middle of that chain. The next step is showing the robots doing useful factory work under ordinary factory constraints.
FAQ
What did Mitsubishi Motors and Highlanders announce?
They signed an MOU to explore joint development of humanoid robots for Mitsubishi manufacturing facilities and possible mass production of Highlanders products at Mitsubishi's Kyoto Plant.
Is Mitsubishi already mass-producing humanoid robots?
No. The confirmed language is exploratory. Mitsubishi says it will examine the feasibility of starting production in early 2027 using currently unused buildings at Kyoto.
Where does the 1,000 units per month number come from?
Electrek and industry coverage reported the target as part of the production plan. It should be read as an ambition until the companies disclose actual line output.
Why is an automaker useful to a humanoid startup?
Automakers understand repeatable manufacturing, quality assurance, safety design, supplier management, and factory operations. Those capabilities are central to turning a humanoid prototype into a product.
Bottom Line
Mitsubishi's Highlanders agreement is not proof that Japan will be shipping thousands of humanoids next year. It is proof that one more automaker sees humanoid robots as an industrial manufacturing problem, not just an AI research problem.
That is the real signal. The humanoid companies that matter in 2027 will not only have better models and better motion. They will have plant partners, deployment loops, quality systems, and enough production discipline to make physical AI repeatable.