Physical AI expansion is pushing Japan's automotive supply chain toward a window for scaling humanoid robots

Overview

The report discusses how advances in Physical AI are moving humanoid robots from prototype validation toward commercialization and whether Japan's automotive industry can evolve from being a robot “user” into a “creator” of complete robots, core components, and software platforms. Goldman Sachs believes that although Japan is a late entrant, labor shortages, automotive supply-chain transformation, mass-production technology, and government support together provide the conditions for catching up.

Core views

Goldman Sachs first substantially raises its long-term outlook for the humanoid robot market: its 2035 sales forecast increases from 1.38 million to 6.48 million units, while the estimated market size rises from approximately ¥6 trillion to US$138 billion, or approximately ¥22 trillion. The upgrade is based on continued improvement in Physical AI capabilities, more manufacturers moving from small-scale prototypes to larger-scale pilot and commercial production, increased availability of open-source tools, and stronger financing and government support. The report also emphasizes the data flywheel: once robots are deployed in real-world settings, they accumulate real-world data, which is then used to train and improve AI software, gradually enabling robots to serve more general-purpose applications. The global competitive landscape is not easy for Japan. Japanese companies account for approximately 30%-40% of the global factory automation market, but companies in the United States, China, and South Korea are building early-stage humanoid robot mass-production systems more quickly, with Chinese manufacturers accounting for most humanoid robot shipments in 2025. The report believes Japan still has an opportunity as a late mover because expanding the range of applications will continue to require innovation in AI capabilities, dexterous hands, and large-scale quality control. Japan ranks fourth globally in industrial robot density, and its social acceptance of robots is second only to South Korea. It also has a complete automotive and auto-parts supply chain that can support the mass production of actuators, reducers, and other hardware. The automotive industry is viewed as the most important early user and a potential manufacturer of humanoid robots. Automobiles are among the world's most complex industrial products manufactured at scale, and automakers possess supply-chain coordination, quality control, and million-unit production capabilities. Goldman Sachs evaluates global automakers along three dimensions—Will, Should, and Could. Will measures whether a company already has mass-production plans or an intention to enter; Should examines the share of the core automotive business and margin trends over the past three years to assess transformation pressure; and Could evaluates execution conditions based on annual production, automotive net cash, and operational capabilities. Annual output above 3 million units, between 1 million and 3 million units, and below 1 million units is assigned 2, 1, and 0 points, respectively. Tesla and Hyundai are currently the most proactive, while Toyota, Honda, and Mitsubishi Motors are relatively better positioned among Japanese automakers. As a robot “user,” Japan faces direct labor constraints. Japan's current working population is approximately 65 million and is expected to continue declining through 2040; manufacturing accounts for approximately 20% of economic activity, while manufacturing, logistics, construction, and nursing care all face labor shortages. Meanwhile, falling robot prices are approaching an economic inflection point where they intersect with labor costs. Goldman Sachs estimates the 2026 manufacturing cost of industrial and research humanoid robots at approximately US$74,000 per unit, while selling prices are expected to decline from US$42,000 in 2025 to US$21,000 in 2035. The robot payback period is equivalent to 2.8 years of labor costs in 2026 and falls further to 1.9 years in 2027. Adoption is expected to proceed in stages rather than through immediate entry into homes and public spaces. The first stage, from 2026 to 2028, covers automotive factories and large logistics centers, where robots primarily perform parts transportation, loading and unloading, and simple picking within fenced or restricted areas. The second stage, from 2028 to 2032, extends to general manufacturing, small and midsize factories, and ordinary warehouses, assuming unit prices decline to US$20,000-US$30,000 and human-robot collaboration safety standards are gradually established. The third stage, beginning in 2032, extends to open environments such as services, nursing care, infrastructure, and homes. At present, only approximately 2% of automotive production processes can directly adopt humanoid robots, mainly sorting and simple logistics. To enter the remaining 98% of complex assembly, wire-harness connection, and visual or abnormal-noise inspection processes, breakthroughs are still needed in AI reasoning, low-latency control, Sim-to-Real, and dexterous hands. The report's manufacturing economics simulation shows that labor costs typically account for 10%-20% of revenue in the automotive and auto-parts industries, with 15% used as the base case. Assuming robot unit prices of ¥3 million-¥9 million and adoption rates of 10%-50%, operating margins could improve by approximately 1-6 percentage points. A more specific example assumes a robot price of ¥4.5 million, system integration costs equal to 50% of the robot price, maintenance costs of 15%, average annual labor costs of ¥8 million, and 250 working days per year. It produces an initial investment of ¥6.75 million, annual net savings of approximately ¥4.245 million, and a 1.6-year payback period. Increasing continuous operating days from 250 toward 365, improving multitasking capabilities, and enhancing quality and yield could all generate additional benefits. As a robot “creator,” Japan's opportunities are more concentrated in mechanical hardware and mass production. Using Atlas's BOM as an example, actuators account for 52%, grippers for 32%, NVIDIA semiconductors for 4%, other sensors for 2%, and batteries for only 1%; this differs markedly from battery electric vehicles, where batteries account for approximately 30%-40% of the BOM. Actuators consist of motors, bearings, encoders, gearboxes, and drives and are highly aligned with traditional auto-parts technologies. The transition to battery electric vehicles reduces the number of automotive parts from approximately 30,000 to approximately 10,000-20,000, giving engine and transmission machining companies an incentive to redirect equipment and capabilities toward robots. The report believes Japan can leverage advantages in precision reducers, lightweighting, dexterous hands, and supply-chain coordination and mass-production quality control similar to the Toyota Production System. Toyota is the potential participant examined most extensively in the report. Goldman Sachs believes Toyota Group meets all four conditions—Needs, Urgency, Technology, and Commitment: it needs to address labor issues; it needs to respond to pressure on the traditional supply chain caused by the battery electric vehicle transition; it has software, hardware, supply-chain, and mass-production capabilities; and management views robots, software-defined vehicles, and AI-driven autonomous driving as areas of non-linear growth. Toyota plans to formally introduce humanoid robots at its Japanese factories beginning around 2028 and subsequently expand them to global sites; in 2026, it was already testing Agility Robotics' Digit at its Woodstock plant in Canada. Toyota Research Institute and Boston Dynamics are also validating Toyota's Large Behavior Model on Atlas. If technological breakthroughs and the mass-production ramp proceed smoothly, Goldman Sachs estimates Toyota could produce approximately 190,000-540,000 humanoid robots in 2035, representing approximately 3%-8% of the global market. The scenario analysis implies robot sales of ¥664 billion-¥1.854 trillion, gross margins of 30%-40%, and gross profit of ¥199 billion-¥742 billion, equivalent to approximately 1.7%-6.2% of Goldman Sachs' estimated total gross profit for FY3/2027; the main text summarizes this as 2%-6% incremental upside. The robot business could also change the valuation structure of automotive companies. Goldman Sachs references the research team's use of 150 times earnings for Tesla's robot business and 60 times for Robotis, starts with the average of 105 times, and discounts it at a 7% WACC to 2035, when Japanese automakers may first generate revenue contributions, resulting in a target P/E of approximately 57 times for the robot business. If robots ultimately account for 5% of Toyota's consolidated revenue, the report's simulation indicates that Toyota's overall P/E could expand by more than 20%; however, this outcome depends on whether the company can translate its strategy into concrete products, customers, and mass-production plans. Government support constitutes another major theme. Japan's Ministry of Economy, Trade and Industry expects the global market for multipurpose robots, including humanoid robots, to reach approximately ¥60 trillion by 2040 and aims for Japanese companies to capture 30%, or approximately ¥20 trillion. The AI robot strategy revised on June 30, 2026 sets targets to deploy 350,000 AI robots by 2030 and approximately 10 million by 2040, prioritizing manufacturing, civil engineering, construction, retail, and security. These sectors account for approximately 40% of Japan's workforce, with jobs suitable for early robotization corresponding to approximately 26 million people. The government plans to provide approximately ¥1 trillion in support over the five years from FY2026 to FY2030, with ¥387.3 billion already allocated in the FY2026 draft budget. Noetra, funded by SoftBank, NEC, Sony, Honda, and others, and the National Institute of Advanced Industrial Science and Technology will develop a domestically produced multimodal foundation model capable of processing language, images, video, sensor, and tactile data. Honda's dexterous-hand project has separately received approximately ¥10 billion in support. At the company level, the report believes Honda's Willow Drive dexterous hand addresses a core bottleneck. It has a maximum fingertip force of 12 kilograms, 20 degrees of freedom, 18 motors, and durability of 8 million cycles, with practical deployment targeted from 2030. However, insufficient disclosure on production volume and costs means it is not yet enough to support earnings or valuation upgrades. Mitsubishi Motors and Highlanders plan to establish monthly capacity of 1,000 units in the second half of 2027, with Mitsubishi responsible for manufacturing and Highlanders for sales, but this remains a future-oriented investment with no near-term revenue contribution. JTEKT could combine its steering, bearing, and factory automation capabilities to develop actuators; under the report's bull-case scenario, related operating profit could reach tens of billions of yen in 2035. Aisin could leverage its experience in transmissions, brakes, metalworking, and durable compact actuators, but it remains at the stage of evaluating market entry. Harmonic Drive has already produced relatively clear commercial data: FY3/26 humanoid robot sales were ¥1.9 billion, FY3/27 guidance is ¥2.3 billion, and it has approximately 15 customers, primarily US startups. First-quarter orders were ¥1.25 billion, and the FY3/29 sales target is ¥7.1 billion. Related sales currently account for approximately 3% of total sales, but their share of quarterly orders has at times reached 5%-10%. Japan-related sales are planned to rise from ¥200 million in FY3/26 to ¥300 million in FY3/27 and ¥3.3 billion in FY3/29. MinebeaMitsumi lists humanoid robots among its five major growth areas, with FY3/35 related sales targets of ¥100 billion in the base case and ¥200 billion in the upside case, including contributions of ¥40 billion, ¥30 billion, ¥20 billion, and ¥10 billion from bearings, motors, analog semiconductors, sensors, and other products, respectively. Renesas's current serviceable market covers approximately 30% of a humanoid robot BOM, with a long-term target of increasing this to 70%. In software and computing, Fujitsu plans to combine its proprietary CPUs, NVIDIA GPUs, AI platform, AI agents, Physical OS, and collaborative control platform. Following the planned launch of AI server production in the first quarter of 2027, the Physical AI business is expected to begin making a full earnings contribution from FY3/28. NEC, as a core Noetra member, will provide video analytics, facial recognition, AI inference, and system integration. Noetra plans to begin construction of computing infrastructure with approximately 27,500 NVIDIA Rubin GPUs in April 2027 and commence operations in June 2028. The valuation approaches and ratings presented for individual companies do not constitute a unified industry target. Toyota and Honda are both rated Buy, with 12-month target prices of ¥3,900 and ¥2,000, respectively, based mainly on the FY3/28E P/B-ROE relationship. Mitsubishi Motors is rated Sell with a target price of ¥330, combining a fundamental value of ¥270 at an 85% weight and a theoretical M&A value of ¥660 at a 15% weight. Aisin is rated Buy with a ¥3,000 target price; Harmonic Drive is rated Buy with a ¥11,000 target price; MinebeaMitsumi is rated Buy with a ¥5,500 target price; Renesas is rated Buy with a ¥5,500 target price; and JTEKT is initiated at Buy with a ¥2,700 target price. NEC and Fujitsu are both rated Buy. The main text's valuation methodology lists target prices of ¥5,190 and ¥4,330, respectively, while the target-price history table separately lists August 21, 2026 target prices of ¥6,000 and ¥4,380; the report does not explain this discrepancy. Goldman Sachs also explicitly identifies three core risks: overemphasizing on-site data and craftsman expertise could cause companies to miss the larger market for simple tasks; reliance on overseas open-source foundation models may become unsustainable due to changes in openness and the digital deficit; and Japan's experience with repetitive tasks in traditional factory automation may not transfer to humanoid robots requiring multipurpose and zero-shot responses. In addition, progress in AI, dexterous hands, safety, lightweighting, and mass-production execution will determine whether the long-term forecasts can be realized.

Analysis framework

The report first reassesses global demand based on Physical AI advances and trends in prototype and commercial production, then evaluates Japan's relative position within the global competitive landscape. It subsequently analyzes labor demand, adoption economics, technological bottlenecks, and supply-chain advantages from the two roles of robot “user” and “creator,” and compares automaker readiness using the Will/Should/Could framework. Finally, it maps industry conclusions to individual companies through Toyota production and earnings scenarios, manufacturing margin simulations, discounting of robot-business P/E multiples, and company-specific target-price methodologies.

Methodology notes

• Industry/Sector Analysis FrameworkSupply-demand framework

  Long-term humanoid robot demand forecast

  The report combines technological progress, mass-production plans, financing, policy support, and actual deployment data to forecast 2035 sales and market size and explains how increased adoption can further improve product capabilities through the data flywheel.

• Competition and Strategy Framework

  Will/Should/Could automaker readiness scoring

  Will measures willingness to enter, Should measures transformation pressure on the core business, and Could measures production, cash, and operational capabilities, enabling comparison of which automakers are more likely to enter the humanoid robot market.

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Robot price, adoption rate, and margin simulation

  The report separately models robot unit prices, system integration and maintenance costs, labor wages, working hours, and adoption rates to estimate payback periods, labor-cost savings, and manufacturing margin improvements.

• Competition and Strategy FrameworkValue chain analysis

  Mapping the humanoid robot BOM to Japanese supply-chain capabilities

  The report breaks down the value shares of actuators, grippers, semiconductors, sensors, and batteries and maps these segments to the motor, bearing, reducer, and mass-production capabilities of Japanese auto-parts companies.

• Valuation MethodPE/PEG valuation

  Discounting the robot-business P/E

  The report references valuation multiples for Tesla's and Robotis's robot businesses and discounts the average P/E of 105 times at a 7% WACC to 2035, when Japanese automakers may generate revenue, deriving a target robot-business P/E of 57 times.

• Valuation MethodPB valuation

  P/B-ROE relationship-based target price

  The fundamental values of Toyota, Honda, Aisin, and Mitsubishi Motors are estimated using the relationship between forecast return on equity and price-to-book ratios, with M&A value added for certain companies.

• Valuation MethodEV/EBITDA valuation

  Enterprise-value multiple-based target price

  Harmonic Drive is valued using FY3/30E EV/EBITDA and industry multiples; JTEKT's fundamental value also uses 5 times EV/EBITDA.

• Valuation Method

  EV/GCI relative to CROCI/WACC and EV/DACF valuation

  MinebeaMitsumi and Renesas determine target prices using approaches combining returns on capital, the cost of capital, and debt-adjusted cash-flow multiples.

• Competition and Strategy Framework

  Three-layer Physical AI architecture

  The report divides Physical AI into a world-model layer responsible for cognition and prediction, a digital-twin layer responsible for simulation training, and an industrial operating-system layer responsible for real-time execution and control, placing Toyota's LBM in the industrial operating-system layer.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Toyota Motor（7203.T）

  The potential beneficiary in complete robots, foundation models, and mass-production systems examined most extensively in the report; Buy, 12-month target price of ¥3,900.

  Strengths

  Combines labor needs, urgency to transform the supply chain, software, hardware, and mass-production technology, and management commitment; potential 2035 output of 190,000-540,000 units.

  Weaknesses

  Robot revenue may not make a meaningful contribution until 2035 at the earliest, and valuation upside depends on the plan becoming concrete and mass production succeeding.

  Comparison

  Relatively ahead of Honda and Mitsubishi Motors among Japanese automakers and viewed by the report as possessing the most complete set of four conditions.

  Risks

  Yen appreciation, EV development delays, battery procurement issues, weaker-than-expected shareholder-return policies, low capacity utilization in China, and delays in robot mass production.

• Honda Motor（7267.T）

  A potential beneficiary of dexterous-hand and humanoid robot R&D; Buy, 12-month target price of ¥2,000.

  Strengths

  Has accumulated expertise from ASIMO development; Willow Drive offers 12-kilogram fingertip force, 20 degrees of freedom, and durability of 8 million cycles and has received government support.

  Weaknesses

  Limited disclosure on future production volumes and cost structure is not yet sufficient to drive earnings forecast or valuation upgrades.

  Comparison

  Identified the dexterous hand as a key humanoid robot bottleneck earlier than most Japanese automakers.

  Risks

  Yen appreciation, sharp increases in raw-material prices, weaker-than-expected sales of new models, insufficient efficiency across multiple EV projects, and declining margins in China.

• Mitsubishi Motors（7211.T）

  Cooperating with Highlanders to produce humanoid robots, with plans to reach 1,000 units per month in the second half of 2027; Sell, 12-month target price of ¥330.

  Strengths

  Possesses automotive manufacturing and mass-production capabilities, and the jointly developed product is intended for customers beyond its internal production lines.

  Weaknesses

  The project is currently an investment for the future and has no near-term revenue contribution.

  Comparison

  Its entry plan is more concrete than those of most Japanese automakers, but it faces competition from US and Chinese manufacturers in AI, hand degrees of freedom, and core supply chains.

  Risks

  Uncertainty surrounding the development of actuator and motor supply chains, mass-produced product competitiveness, and achievement of 2027 capacity.

• JTEKT（6473.T）

  A potential beneficiary in actuators, steering, bearings, and factory automation; initiated with a Buy rating and a 12-month target price of ¥2,700.

  Strengths

  Possesses three core capabilities in steering, bearings, and FA and may participate in Toyota Group robot projects; related operating profit could reach tens of billions of yen in 2035 under the bull-case scenario.

  Weaknesses

  No clear commercialization timeline has been established.

  Comparison

  South Korea's HL Mando and Europe's Schaeffler have already commercialized humanoid robot actuators, while JTEKT remains in the development stage.

  Risks

  Lower Toyota production, yen appreciation, and further delays in Toyota's humanoid robot development.

• Aisin（7259.T）

  A potential beneficiary of highly durable compact actuators and metalworking capabilities; Buy, 12-month target price of ¥3,000.

  Strengths

  Experience in transmissions, brakes, rear-wheel steering, and aluminum processing offers strong technological synergies with robot actuators.

  Weaknesses

  Management is still evaluating whether to enter the robot business.

  Comparison

  Like JTEKT, it is a potential core component participant within Toyota Group, but its commercialization readiness is at an earlier stage.

  Risks

  Higher aluminum prices, continued Middle East impacts, and price negotiations with major customers.

• Harmonic Drive Systems（6324.T）

  A direct beneficiary in precision reducers; Buy, 12-month target price of ¥11,000.

  Strengths

  Has already generated actual orders and sales, with approximately 15 customers; targets humanoid robot sales of ¥7.1 billion in FY3/29, and AI robots are expected to become an earnings growth driver from FY3/28.

  Weaknesses

  Related sales currently account for only approximately 3% of total sales, revenue recognition is relatively slow, and methods for increasing capacity will need to improve as the market expands.

  Comparison

  Described by the report as a representative humanoid robot stock in Japan, with clearer commercialization data than other Japanese supply-chain companies.

  Risks

  Weaker investment appetite among technology manufacturers and EMS companies, lower mass-production investment by industrial robot and semiconductor equipment manufacturers, and greater allocation of resources to the auto-parts business.

• NEC（6701.T）

  Benefits through Noetra's domestically produced multimodal foundation model and system-integration business; Buy.

  Strengths

  Can provide video analytics, facial recognition, AI inference, and cross-industry implementation capabilities and is participating in the construction of computing infrastructure using approximately 27,500 Rubin GPUs.

  Weaknesses

  The earnings contribution depends on the progress of Noetra's infrastructure construction, model development, and cross-industry deployment.

  Comparison

  More focused on foundation-model implementation and system integration; the main valuation-methodology text lists a target price of ¥5,190, while the historical table separately lists an August 21, 2026 target price of ¥6,000.

  Risks

  Declining corporate appetite for IT investment, weak capital expenditure by telecommunications operators, replacement by overseas products, and the emergence of loss-making projects.

• Fujitsu（6702.T）

  A potential beneficiary of AI servers, Physical OS, and robot collaborative-control platforms; Buy.

  Strengths

  Can combine proprietary CPUs, NVIDIA GPUs, an AI platform, and partner devices and plans to promote standardization and open-sourcing of its collaborative-control platform.

  Weaknesses

  The business is not expected to make a full earnings contribution until FY3/28 and depends on server production and joint development with partners.

  Comparison

  More focused on computing infrastructure and control platforms than NEC; the main valuation-methodology text lists a target price of ¥4,330, while the historical table separately lists an August 21, 2026 target price of ¥4,380.

  Risks

  Declining IT services or hardware demand due to economic deterioration, weak telecommunications capital expenditure, and the emergence of loss-making projects.

• MinebeaMitsumi（6479.T）

  A diversified beneficiary across bearings, motors, analog semiconductors, sensors, and dexterous-hand components; Buy, 12-month target price of ¥5,500.

  Strengths

  Its product portfolio closely overlaps with humanoid robots, with FY3/35 related-sales targets of ¥100 billion in the base case and ¥200 billion in the upside case, and it is already jointly developing robot hands with three companies.

  Weaknesses

  Achievement of the long-term targets depends on customer mass production, cross-product integration, and market expansion.

  Comparison

  Can cover more robot-component categories than a single-product reducer manufacturer.

  Risks

  Larger-than-expected increases in investment costs, declining automobile production, and exchange-rate fluctuations.

• Renesas Electronics（6723.T）

  A potential beneficiary of semiconductors for robot computing, perception, motor control, and power management; Buy, 12-month target price of ¥5,500.

  Strengths

  Its current products can cover approximately 30% of a humanoid robot BOM, with a long-term target of 70%; automotive technology expertise and the Renesas 365 platform can support coordinated hardware-software design.

  Weaknesses

  The humanoid robot market remains in a stage of rapid technological change, high product variety, and low production volumes.

  Comparison

  Compared with a single-chip supplier, the report emphasizes the combination of its hardware portfolio with cloud-based design, optimization, and lifecycle-management capabilities.

  Risks

  Slower recovery in consumer-electronics demand, stagnant automobile production, prolonged semiconductor inventory adjustments, delayed recovery in industrial-equipment demand, and yen appreciation.

Key data

• 2035 humanoid robot sales forecast6.48 million unitsPrevious forecast was 1.38 million units
• 2035 humanoid robot market sizeUS$138 billion, approximately ¥22 trillionPrevious forecast was approximately ¥6 trillion
• Japanese companies' global factory automation share30%-40%Contrasts with their limited presence in complete humanoid robots
• Toyota's potential 2035 productionApproximately 190,000 to 540,000 unitsCorresponding to a global share of approximately 3%-8%
• Toyota robot-business gross profit contribution¥199 billion to ¥742 billionEquivalent to approximately 1.7%-6.2% of estimated total gross profit for FY3/2027, summarized in the main text as 2%-6% upside
• 2026 humanoid robot manufacturing costApproximately US$74,000/unitBOM estimate for industrial and research humanoid robots
• Humanoid robot price forecastDeclining from US$42,000 in 2025 to US$21,000 in 2035Mass-production effects drive price declines
• Robot payback period2.8 years in 2026, 1.9 years in 2027Measured against equivalent labor costs
• Automotive processes currently suitable for robotizationApproximately 2%The remaining 98% of complex processes require AI and dexterous-hand innovation
• Manufacturing operating-margin upsideApproximately 1-6 percentage pointsSimulation results assuming robot prices of ¥3 million-¥9 million and adoption rates of 10%-50%
• Key Atlas BOM sharesActuators 52%, grippers 32%, semiconductors 4%, other sensors 2%, batteries 1%Indicates mechanical hardware is the main value segment
• Japan's AI robot deployment targets350,000 units by 2030, approximately 10 million by 2040AI robot strategy revised in 2026
• Japanese government supportApproximately ¥1 trillion for FY2026-FY2030The FY2026 draft budget allocates ¥387.3 billion
• Honda Willow Drive dexterous hand12-kilogram fingertip force, 20 degrees of freedom, 18 motors, durability of 8 million cyclesPractical deployment targeted from 2030
• Mitsubishi Motors' planned capacity1,000 units per monthPlanned for establishment in the second half of 2027 in cooperation with Highlanders
• Harmonic Drive FY3/29 targetHumanoid robot sales of ¥7.1 billionFY3/26 actual sales of ¥1.9 billion and FY3/27 guidance of ¥2.3 billion
• MinebeaMitsumi FY3/35 targetBase case ¥100 billion, upside case ¥200 billionCovering bearings, motors, analog semiconductors, sensors, and other products
• Renesas BOM coverage targetIncrease from approximately 30% to 70% over the long termCovering brain and motion, perception, actuation and motor control, and power management
• Noetra computing infrastructureApproximately 27,500 NVIDIA Rubin GPUsConstruction planned to begin in April 2027, with operations starting in June 2028

Impact & implications

The report believes humanoid robots will first change the labor and cost structures of automotive and auto-parts manufacturing before spreading to general manufacturing, logistics, services, and homes. Opportunities for Japanese companies are not limited to complete robots and may be more significant in actuators, precision reducers, bearings, motors, dexterous hands, semiconductors, foundation models, and system integration. If companies can establish concrete mass-production, customer, and earnings plans, robot businesses could both contribute profit and raise valuation multiples for automotive and auto-parts companies; conversely, technological and execution delays would make long-term scenarios difficult to realize.

Risks

• Overemphasizing on-site data and craftsman expertise could cause development resources to overlook the larger market for simple tasks.
• Reliance on overseas open-source foundation models could become an unsustainable business model if models cease to be open or the digital deficit expands.
• Japan's successful experience in traditional factory automation may not apply to humanoid robots requiring multipurpose and zero-shot responses.
• At present, only approximately 2% of automotive processes can directly adopt humanoid robots; the remaining complex processes depend on breakthroughs in AI, Sim-to-Real, low-latency control, and dexterous hands.
• Robot falls could injure personnel and damage equipment, while weight also increases energy consumption and shortens operating time; safety and lightweighting remain constraints.
• Long-term earnings and valuation scenarios depend heavily on product prices, performance, mass-production ramp-up, customer adoption rates, and execution of commercial plans.
• Yen appreciation, declining automobile production, rising raw-material costs, EV project delays, and slower corporate IT or equipment investment represent specific downside risks for the relevant companies.

What to watch

• Monitor the speed at which humanoid robots move from large-scale prototypes to commercial production and whether the forecast of 6.48 million units of demand in 2035 gains order support.
• Track whether robot prices, labor costs, and payback periods approach the expected economic inflection point.
• Watch for technological breakthroughs in AI reasoning, Sim-to-Real, low-latency control, and dexterous hands with more than 20 degrees of freedom.
• Track Toyota's concrete plans to formally introduce robots at Japanese factories around 2028 and subsequently roll them out to global sites.
• Monitor Honda's disclosure of Willow Drive's mass-production scale, cost structure, factory validation results, and commercialization progress toward 2030.
• Observe whether Mitsubishi Motors and Highlanders can deliver monthly capacity of 1,000 units in the second half of 2027 and establish actuator and motor supply chains.
• Track whether orders convert into sales after Harmonic Drive's major customers expand production beginning in 2027 and whether Japan-related sales can reach the FY3/29 plan of ¥3.3 billion.
• Monitor Fujitsu's FY3/28 earnings contribution, Noetra's computing infrastructure construction in 2027-2028, and NEC's system-integration deployments.
• Track the implementation of the Japanese government's approximately ¥1 trillion support plan and its deployment targets of 350,000 units by 2030 and 10 million units by 2040.
• Monitor whether Japanese companies can expand from strengths in actuators, reducers, and bearings into complete robots, foundation models, and industrial operating systems.