Robot disruption is getting closer with humanoid robots
There has been a lot of attention on recent developments in artificial intelligence (AI), with new releases like GPT-5 and video AI systems like Sora and Nano banana. While all this has been going on there has also been rapid development of ‘embodied AI’, which adds AI to robots to make them more flexible and capable, and in particular to make humanoid robots that interact, are dexterous, and agile. National strategies like China’s Intelligent Manufacturing 2025, Japan’s robotics initiatives, Germany’s Industry 4.0, and U.S. innovation policies all actively promote the robot industry.
General purpose robots include quadrupeds, wheeled robots, and humanoids, as in the figure above. Some have a base with human arms and torsos, because the legs and waist of a humanoid are expensive. However, there are many new two legged humanoid robots intended for logistics, industrial assembly and automation tasks. Humanoid robots operate in an environment built for people, able to perform physical tasks in unstructured spaces with minimal reprogramming.
Figure 1. US Humanoid robots
Note: Sanctuary AI is Canadian. Amazon has begun testing Digit at their R&D site. Hyundai owns Boston Dynamics and will use Electric Atlas in its Georgia US car plant
This post looks at the state of play of industry use of robotics and AI with a focus on embodied AI and humanoids, starting with recent reports from McKinsey on automation and robotics. The 2025 International Federation of Robotics annual report on sales and use and a recent position paper on humanoids is covered, followed by a report on Chinese robotics shows in Beijing and Shanghai. China’s humanoid robotic strategy is discussed, then Australia’s robotics strategy and roadmap. There is a section on construction robots with a new McKinsey report on humanoids in construction, and discussion of the Robot as a Service model and the potential of humanoid subcontracting in construction. The Amazon robotics system used in their newest warehouses is described. This post covers a lot of ground and is longer than usual.
McKinsey’s Robotics Research
A June 2025 McKinsey report Will embodied AI create robotic coworkers? opens with ‘General-purpose robotics funding grew fivefold from 2022 to 2024, surpassing $1 billion in annual investment, with leading start-ups such as Figure AI, Skild AI, and Agility Robotics raising hundreds of millions of dollars. Patent filings have also surged, with a 40 percent CAGR in volume since 2022…. China has designated embodied AI a national priority, anchoring a $138 billion innovation fund’ [1]. McKinsey estimated that the general-purpose robotics market could reach about $370 billion by 2040.
The report found general purpose robots capable of grasping, lifting and placing items are not here yet, ‘but the building blocks are emerging fast.’ Barriers like task specific data and battery life will be addressed, supply chains for components and sensors will develop, and autonomy and dexterity will improve. The report had the figure below on industry tasks.
Figure 2. Robots and tasks
There was a second McKinsey report in July, A Leap in Automation: The New technology Behind General-Purpose Robots. This starts with ‘advances in software are now making general-purpose robots viable by enabling them to learn, adapt, and act in real time, without human intervention. Simultaneously, hardware improvements are optimizing robot dexterity, sensing, and power.’ Foundation models provide the ‘most straightforward path to improve robot capabilities’ to skill levels comparable or greater than humans. That report concludes ‘software and hardware challenges ahead may seem daunting, but recent advances suggest that general-purpose robots could take workplace automation to new heights. The only questions are how quickly progress will occur and whether companies will embrace change rapidly or hesitate.’
Figure 3. Robotic capabilities and foundation models
Also in July, the McKinsey 2025 Technology Trends Outlook section on robotics noted that interest and innovation metrics for robotics increased by double digits from 2023 to 2024 and the market is expanding ‘with adoption in emerging sectors like agriculture and healthcare, the development of cobots, and broader deployment in manufacturing and logistics.’ (Cobots are collaborative robots for shared workspaces).
There were two examples of robotic foundation models: Covariant’s RFM-1 helps robots understand how objects move and interact in the real world, follow language-based instructions, and reflect on their own actions; and Figure AI’s Helix, a vision-language-action model that enables humanoid robots to perform complex tasks, enabling dynamic object recognition and collaboration without prior training.
The Trends Outlook also had four examples of industrial applications: Amazon’s AI-driven cobots in its warehouses automate pick-and-place and palletization tasks; Tortuga AgTech’s robots identify and pick fruit; KFoodtech’s BotBob cooks traditional Korean stew; and Sanctuary AI’s advanced haptics tactile sensors and Meta AI’s Digit 360 have robotic dexterity. Two humanoid robot examples were: GXO Logistics agreement with Agility Robotics to deploy the Digit robot to automate tasks like moving totes and managing palletization; and Hyundai plans to deploy their Boston Dynamics Electric Atlas in automotive manufacturing as a cobot.
McKinsey claims the market for service robots in areas such as logistics, hospitality, and agriculture has been growing at 20 to 35 percent annually, and many of these robots do not resemble humans. For example, autonomous mobile robots (AMRs) that transport medical supplies in hospitals or materials within a warehouse resemble a robotic household vacuum more than a person.
Figure 4. European and UK humanoid robots
Note: NVIDIA announced a partnership with 1X in March to develop home robots. Talos II is a heavy lift robot, Ameca a companion robot, and Kime operates a drinks stand.
International Federation of Robotics Annual Report
The International Federation of Robotics (IFR) is a nonprofit industry trade group. The membership in Figure 5 is mainly European, American and Japanese companies, and the absence of many Chinese manufacturers is striking (e.g. non-members include Siasun, Step and HGZN in industrial automation, and Unitree, UBTech, Hanson, and Hikrobot in service robots). The many Chinese R&D institutes are also conspicuous by their absence.
Figure 5. IFR members
Source. Note: In October 2025 ABB sold its robotics business to Softbank.
Each year IFR issues a World Robotics Report, which has a press release with some figures on trends. The 2025 report came out in September with data for 2024, and is divided into two sections, one on industrial robots and the other on service robots. Industrial robots are ‘automatically controlled, reprogrammable multipurpose manipulator programmable in three or more axes’ and service robots are a ‘robot in personal use or professional use that performs useful tasks for humans or equipment,’ with the key difference that an industrial robot makes a physical product in a factory. Drones and autonomous vehicles are excluded. The IFR follows the International Standards Organisation definitions of robots.
Industrial Robots
KUKA (German), ABB (Swiss), Fanuc and Yaskawa (Japan) dominate industrial robot manufacturing with about 57% of the market. Other major manufacturers are Mitsubishi, Denso and Nachi (Japan), Omron and Standard Bots (US) and Staubli (Swiss).
IFR members reported a total number of industrial robots in operational use of 4,664,000, an increase of 9% on 2023. For many years the use of industrial robots was primarily in the two industries of electronics and cars but, while those are still the two biggest users, they now account for less than half of the market as the Metal and machinery and Other industries categories increase. In 2014 Automotive was 43% of the robot market, down to 23% in 2024.
Figure 6. Industry use of industrial robots
For the last four years, annual installations were over 500,000 units, and in 2024 Asia accounted for 74% Europe 16%, and the Americas 9%. IFR forecasts annual installations rising from 575,000 in 2025 to 708,000 in 2028, with 550,00 of the 2028 installations in Asia
China is the largest market, with 54% of global installations and 295,000 industrial robots installed in 2024. Chinese manufacturers domestic market share has increased to 57%, up from 26% a decade ago, and China’s operational robot stock of over 2 million is the largest of any country. Japan is the second-largest market with 44,500 units installed in 2024 and 450,500 units in use. The US installed 32,200 and had 393,700 robots, South Korea installed 30,600 and had 392,900 robots, and Germany installed 27,000 and had 278,900 robots. These are by far the five largest markets.
Service Robots
The IFR category of service, mobile, and medical robots includes robots that can autonomously drive around warehouses and pick items off shelves, robots for professional cleaning, search and rescue robots, and robots that can conduct laboratory tests or assist with surgery. There is much less data on service robots and the IFR notes: ‘All numbers based on a sample of 293 producers of total 944 manufacturers. Data is not projected to the whole industry. Sample composition varies each year.’ This recognises the absence of major robot developers and users like Amazon as members.
Figure 7. Service robot installations 2024
Note: Search, rescue and security robot installations were 3,128.
The IFR explains the ‘service robot industry is more heterogenous than the industrial robot industry’. There are 944 service robot producers, excluding prototyping services and system integrators, many companies are in the funding or prototyping stage and 80% are small or medium-sized enterprises with up to 500 employees.
In 2024, sales of professional service robots were over 199,000 units, and grew by 9% from 2023, and for consumer use 20.1 million units were sold, an increase of 11%. Sales of medical robots in 2024 were particularly strong with close to 16,700 units sold and a growth rate of 91%.
Humanoid Robots
The final page of the IFR press release is on humanoid robots, with four bullet points:
· There is no massive use today but serial production in preparation;
· Manufacturers are building humanoid robots for R&D purposes for several years and producing on demand;
· New companies in the market do build humanoids at a demonstrator or prototype stadium for first trial applications;
· Application fields of humanoids still have to be determined and proven in practice.
Despite the garbled syntax, the IFR believes humanoid robots are not yet happening, because they are still in the R&D and prototype stage. There is also no section on humanoid robots on their web site. In August 2025 the IFR published a position paper on Humanoid Robots – Vision and Reality. That looked at regional differences: the US has a focus on practical applications; China is using humanoids in service sectors, and the strategy is to establish a supply chain for key components that is scalable; in Japan robots are regarded as companions rather than tools; and in Europe the focus is on ethics, human-centric design, and the social impact of robots. The paper concluded with ‘human-like dexterity and adaptability, humanoids are well placed to automate complex tasks with which current robots struggle using traditional programming methods. However, mass adoption as universal household helpers may not happen in the near or medium term.’
The narrow IFR/ISO definitions of industrial and service robots are becoming unworkable as autonomy increases and the range of robot forms widens to include AMRs, various drones, quadrupeds, and of course humanoids. While some of these are commonplace, like Amazon bots and autonomous mobile robots in warehouses, and others like autonomous drones for inspections and cleaning are becoming more widely used.
Robots in China
There is an industry expo or event on automation and robotics most months of the year in China, and they are huge. The World Robot Conference 2025 was in Beijing in August. The five day event had over 1,500 exhibits from more than 200 companies, both Chinese and international, and featured ‘embodied AI’ in a wide variety of robots. There is a 20 minute video on YouTube with bird and quadruped robots, humanoid robots dancing, playing music, pouring drinks, greeting people, folding laundry, preparing and delivering meals, as well as the more familiar pick and place, fetch and carry, maintenance, monitoring and industrial robots. US company Skild demonstrated their ‘shared, general-purpose brain’, an AI that can learn to perform tasks across different robot forms. They intend to operate a business model similar to OpenAI, selling access to a foundation model for customers to create robots for different tasks in a wide range of scenarios.
Then in September there was the Robotics Show in Shanghai. Held annually since 2012, it covered everything from industrial and humanoid robots to logistics solutions, including collaborative robots (cobots), service robots, robotics for palletizing and material handling, and mobile systems. Exhibitors also included core component suppliers like servo motors, controllers, gear units and machine vision. This year’s event saw the launch of thousands of new products.
Figure 8. Chinese humanoid robots
In a report on the Shanghai Robotics Show, Foundamental partner Patric Hellerman says: ‘China's domestic robotics market provides manufacturers with a testing ground and scaling platform that no Western market can replicate, creating an insurmountable advantage in deployment experience and product iteration speed. For anyone building upstream robotics hardware (arms, cobots, electrical or mechanical components) in the West, this creates an uncomfortable business reality. Your ONE domestic market isn't large enough to provide equivalent learning opportunities. Your ONE customer base isn't dense enough to support the same rate of iteration. Your ONE ecosystem doesn't have the same depth of suppliers, integrators, and supporting infrastructure.’
Chinese manufacturers are selling thousands of units monthly to domestic customers, and intense domestic competition forces low prices, so Chinese prices are three to six times lower than Western market prices. A ‘thirty-thousand-dollar robot in Shanghai becomes ninety thousand to one hundred eighty thousand dollars in Munich or Manchester.’ Western competitors cannot match those prices ‘without destroying their business models entirely.’ Less intense Western competition allows market segmentation and premium pricing.
Hellerman argues China is following the same strategy used to build dominance in solar panels, batteries and EVs: ‘Western robotics companies burning venture capital to build factories are bringing equity tools to a debt fight, creating unsustainable cost structures that cannot compete with manufacturers accessing patient, cheap capital designed specifically for industrial buildout.’ Western venture capital is ‘structurally disadvantaged for capital-intensive hardware manufacturing competing against systematic debt availability, subsidies, and export incentives.’
Figure 9. More Chinese humanoids
In December there will be the Shanghai Humanoid Robot and Robotics Technology Expo, with ‘robot core components, full machine manufacturing, and system integration solutions’ the expo aims to accelerate innovation and commercialisation in humanoid robotics. ‘Backed by Beijing’s strategic support and deep technological resources, the expo promotes industrial scaling and collaborative application development across scientific, industrial, and service domains.’
Figure 10. Humanoid robots new in 2024
China’s Robot Strategy
China has gone from being a technological follower a decade ago to an innovator in industrial automation and robotics, with advances in control systems, sensors, manipulators and drive mechanisms. Its robot industry has grown through a combination of research funding, technological breakthroughs, market diversification, and a great deal of policy support.The rapid growth of the country's robotics industry has robots now used in warehouses, education, entertainment, cleaning services, security inspection and medical facilities. A supply chain of components is now established.
China is prioritizing AI development in robotics and embodied intelligence. The 14th Five-Year Plan for the Development of the Robot Industry by 2025 was released in in 2021, followed by the Robot+Application Action Plan in 2023. In April 2025 the Ministry of Industry and Information Technology claimed the country holds over 190,000 active robotics patents, accounting for roughly two-thirds of the global total. The AI+ plan announced in August set adoption targets above 70% by 2027 and 90% by 2030, and an intelligent economy and society by 2035. The central government included embodied intelligence as one of four future industries in its 2025 work report: ‘We will establish a mechanism to increase funding for industries of the future and foster industries such as biomanufacturing, quantum technology, embodied AI, and 6G technology’ (p.18).
A roadmap to 2030 adopted at the second Embodied AI Conference in March 2025 had three stages:
1. 2025 to 2027, foundation: Shared datasets and open middleware.
2. 2028 to 2030, scaling: Deployment in factories, logistics, elder‑care pilots
3. Post‑2030, generalization: Mass‑market humanoids and composite robots.
These stages are in progress. The Huisi Kaiwu platform was unveiled in March as a universal software system for robotic intelligence, allowing robots to be programmed for different tasks. The project is overseen by the Beijing Innovation Center of Humanoid Robots, jointly funded by the Ministry of Industry and Information Technology, the Beijing municipal government, private firms and research institutions. A standard for robotic elder care was released in February, with technical benchmarks for product design, manufacturing, testing, and certification. This year will see mass production of humanoid robots as UBTech Robotics secured orders for more than 500 from car companies, expecting to reach 1,000-2,000 orders by the end of 2025, Zhiyuan Robotics plans to expand sales to 3,000-5,000 by the end of 2025, and Unitree has sold 1,000 humanoid robots.
A major problem for humanoid robots is the lack of training data because, unlike large language models that have been trained on the internet, they require data on work processes. To tackle this data deficiency AgiBot set up a data collection factory in Shanghai in 2024, using around 100 robots to gather 30,000 to 50,000 data points a day. At the end of the year AgiBot released the AgiBotWorld dataset, with ‘more than 1 million trajectories from 217 tasks across five major scenarios’ as open source training data.
The Mercator Institute for China Studies 2023 report on China’s Innovation Chain found 173 High-Tech Industrial Development Zones (HIDZ). These are under the Ministry of Science and Technology and host 84% of State Key Laboratories. There are also 230 Economic and Technological Development Zones (ETDZs), which are similar to HIDZs but under the Ministry of Commerce. The report uses the Shenyang robotics and smart manufacturing cluster as an example. It has the State Key Laboratory for Robotics at the Shenyang Institute of Automation, and the anchor tenant is Siasun, a robotics company spun out of the Chinese Academy of Sciences (a government research agency) with subsidiaries in Singapore, Thailand, Malaysia, and Germany.
The Chinese government has established other innovation hubs focused on robotics supported by the national and provincial governments, with incentives for manufacturers to locate near the research centres. Dongguan robot city has a robotics research institute at the core, and in Shenzhen the Huawei (Shenzhen) Global Embodied Intelligence Industry Innovation Center started in 2024, partnering with local robotics firms Leju Robotics, Zhaowei Electromechanical, and Daju Robotics.
The Mercator Institute report concluded: ‘China’s central government is focusing on extracting more economic value out of research spending and developing market-ready technologies that may provide an advantage in a geopolitical era where technological supremacy is seen as a key source of power. This desire to commercialize research has led China’s government to reorganize its innovation system around the idea of the innovation chain. To achieve this, Beijing has embarked on a long and painful process of reforming funding programs, institutional support, and zoning policies….Additional policies and updates to evaluation criteria for scientific research are also shifting emphasis to marketization over other R&D outputs.’
Figure 11. Humanoid robots forecast
Australia’s Robotics Strategy
Australia's May 2024 robotics strategy is a ‘framework for a robotics and automation ecosystem to increase productivity and competitiveness.’ The strategy has four themes: national capability; increasing adoption; trust, inclusion, and responsible development; and skills and diversity. The strategy seeks to ‘use robotics and automation to help meet challenges like climate change, an aging population, and geopolitical risks.’
The strategy has as examples autonomous trucks in mining (700 in 2022), SwarmFarm Robotics in agriculture, and Finisar Australia manufacturing optical switching technology. It argues Australia has research expertise and growing local capability. However, a year and a half after the strategy was launched, its concluding claim is either aspirational or delusional: ‘Australia has unique advantages in contributing to the AI systems that will continue to revolutionise robotics into the future. Not only is Australia a leader in AI R&D, but our companies and researchers are already actively developing robotics software solutions to address problems across a range of industries.’ While mining and agricultural systems are world class, the lack of funding for R&D and startups means Australia is far from the technological frontier.
Robotics Australia Group Roadmap
The Robotics Australia Group (RAG) has over 70 members across 14 industry chapters. There was a 2025 update of their Roadmap, which said Australian companies are ‘starved’ of capital compared to international competitors, as over the last nine years 10 companies have raised only USD$255 million, and another 144 companies shared USD$30m [see 1]. The Roadmap has many examples of development and use of robots in Australian industries like autonomous machinery in mining, palletizing and lifting in manufacturing, and picking and packing in agriculture. There is a chapter on construction robots, with these examples of applications in Australia:
· Autonomous or semi-autonomous earth-moving machinery;
· Brokk remote controlled demolition robot;
· Autonomous or semi-autonomous robotic cranes;
· Modbotics manufactures timber house frames with a Randeck Robotics system;
· Schindler’s RISE robot used for drilling in elevator shafts;
· Robotics 3D concrete printing;
· Aptella uses HP SitePrint for autonomous road marking and site layout;
· Inspection and monitoring drones using Presien AI vision technology;
· FlyFreely drone management platform;
· FBR’s Hadrian X bricklaying robot, now available as a subcontract.
Construction Robots
A post in August last year included 17 robots in use onsite: three for bricklaying, five for layout printing and surveying, four for reality capture and site monitoring, and one for each of drywall finishing, anchor hole drilling, rebar fixing, underfloor insulation, and solar farm construction. None were humanoids. Following the ISO definition used by IFR, excluded were exoskeletons, offsite manufacturing, remote controlled drones, autonomous excavators and graders, and onsite 3D concrete printing (which was the subject of this 2022 post).
Last week McKinsey published Humanoid Robots in the Construction Industry: A Future Vision (Oct. 17, 2025), and comparing that to their reports from a few months ago at the beginning of this post shows how fast the technology is developing. ‘To determine what construction tasks humanoids could perform in ten or more years, we assessed activities based on the level of collaboration involved, ease of training, dexterity requirements, and the structure of the work environment. We then categorized use cases according to the degree of relative task feasibility’, in Figure 12 below.
Figure 12. McKinsey on construction tasks and humanoids
McKinsey thinks construction leaders should not wait for full-scale deployment to begin preparing for a future in which humanoids and humans work together. Although humanoids are ‘a vision for the future workplace, rather than an immediate reality’, they are ‘expected to demonstrate high commercial feasibility in large-scale infrastructure projects and in residential, commercial, or institutional construction.’ The Bain and Company 2025 Technology Report has a section called Humanoid Robots: From Demos to Deployment that made the same argument on preparation, and forecast they will be used in construction in 10 years for onsite materials handling when battery life has been extended.
It is worth noting some other examples of robots already being used on construction sites. First, robotic excavators. These systems are installed on excavators to allow autonomous operation: Built Robotics (US); Bedrock Robotics (US, raised $80m in July); and Gravis Robotics (Swiss). Second, equipment with remote operation: Flywheel AI (US, excavators); Brokk (Sweden) demolition robot; and Nyro (US), a humanoid for demolition. Although only one of these examples is a humanoid, Nyro shows how humanoids are likely to make their way onto construction sites, as remote control allows the ‘brain’ to learn by doing, the same way as apprentices in construction have for centuries. Once ‘training’ is complete, autonomous operation follows, which is how robotic excavators have been developed.
The business model in construction could be ‘Robot as a service’ (RaaS), where businesses lease robotic devices to gain the benefits of automation, such as lower costs, flexibility, and scalability, without the large initial investment and maintenance associated with ownership. This would allow the industry’s many small and medium size contractors and subcontractors to use robotic equipment. With leased humanoids trained for materials handling or specific tasks, there may be robotics subcontractors where the manufacturer is also the subcontractor, like FBR is doing with their truck mounted Hadrian X bricklaying robot.
Amazon’s Robotic System
Amazon has been developing robots for use in their warehouses since taking over Kiva Systems in 2012, and now has deployed more than 1 million robots. In 2024 a new fulfillment centre in Shreveport Louisiana uses eight robotic systems, plus two AI models to manage the robots and workflow, one a foundation model the other an agentic model. An October post followed a package through the 10 robots ‘that are supporting the next generation of package fulfillment at Amazon’:
1. Sequoia uses AI, robotics, and computer vision to consolidate inventory and free up storage.
2. Hercules is a drive unit that finds and brings pods of items to employees.
3. Titan is another drive unit that brings items to employees, it can lift twice as much as Hercules (up to 2,500 pounds, over 1,100 kilos). Titan and Hercules are confined to areas where only authorized robotic specialists can enter, and read barcodes that are stickered to the floor as navigation coordinates.
4. Vulcan has a sense of touch, and picks and stows items from the highest and lowest inventory pods so employees don’t have to climb ladders or crouch.
5. Blue Jay is a ceiling-mounted robotic system that coordinates multiple robotic arms that simultaneously pick, stow, and consolidate packages.
6. Sparrow is a robotic arm that picks up and moves individual items from containers into specific totes.
7. Packaging station uses sensors to measure an order’s dimensions and creates a correctly sized, protective bag.
8. Robin is a robotic arm that grabs packages from conveyor belts and puts them onto robotic drive units to be moved.
9. Cardinal is similar to Robin, a robotic arm that uses advanced AI and computer vision to select one package from a pile, lift it with air suction, read the label, and place it in the appropriate cart.
10. Proteus works with Cardinal to move carts from the outbound dock area to the loading dock. Proteus is Amazon’s first fully autonomous mobile robot, using sensors to detect and avoid objects.
Conclusion
Robotic technology is rapidly advancing, and in environments like factories and warehouses the variety and use of robots is increasing quickly. For many tasks, non-humanoid automation is more practical, such as a fixed robot arm for welding or picking, and drive units or wheeled couriers for deliveries. Amazon is a good example with their AI coordinated robotic systems, and many firms will follow Amazon’s lead with automated stowing, picking, moving and packing items. However, the human form has the great advantage of functioning in an environment built for humans, therefore companies will keep investing in development of humanoid platforms.
A year ago humanoid robots were still at the experimental and prototyping stage. Now, however, humanoid robots are breaking out, and some are already being used in accommodation, food services, manufacturing, transportation and warehousing. Although these are early days and many humanoid uses are more like trials than mass deployments, production is increasing quickly. The foundation models for humanoids being trained by Field AI and Figure AI will have construction industry applications.
Figure 13 below is from a recent Bain and Company Brief that said ‘we anticipate steady growth in the deployment of humanoid robots through 2030, followed by a rapid uptick thereafter. Market projections range from $38 billion to more than $200 billion by 2035. Global funding in humanoid robotics start-ups is growing quickly, up from about $308 million in 2020 to $1.1 billion in 2024 [US dollars]. Several global tech leaders along with specialized robotics start-ups are ramping up production.’
Figure 13. Humanoid production plans
The Chinese model of state-led development, long-term planning, supply chain and infrastructure investment, combined with selective integration into global markets, is working for humanoids as it did for solar panels, batteries and EVs. Whether one admires this model or not, it has been a success and has delivered technological progress. China is now restructuring and integrating the innovation chain, aiming at technological self-reliance and developing a supply chain of components for humanoids like servo motors, controllers, gear units and machine vision. Although at present Chinese robots may not be as sophisticated as some others, they are much cheaper.
Humanoid robots are now taking on tasks in controlled environments like car plants and warehouses, where they could quickly become widely used for repetitive tasks. In other industries there will be a period of human/robot collaboration, as robots take on simple tasks, and dull, dirty, or dangerous work. Then, as humanoid brains based on foundation models develop, and their strength, dexterity, and battery life improves, more applications will follow. Many of the jobs in structured environments that are repetitive and tightly controlled could be done by humanoids in the near future (say five years), and in ten years there will probably be humanoids doing some but not all of the work in most workplaces, including construction sites.
[1] Humanoid 2025 funding in USD includes:
· Agility Robotics (US) $400m
· Apptronik (US) $403m, has partnerships with Amazon, Mercedes-Benz, Walmart and Jabil to test Apollo in warehouses and automotive plants
· Field AI (US) raised $400m for ‘Field Foundation Models’ (general purpose AI models)
· Figure AI (US) over USD$1 billion raised in September valued the firm at $39 billion, has a multiyear agreement to deploy its robots in BMW’s South Carolina plant
· Fourier (China) $109m
· Neura (Germany) $136m
· Persona (US) $27m
· UBTech (China) listed on Hong Kong stock exchange in 2023, raised $119m in 2025
· Wandercraft (France) $75m
Subscribe on Substack here
https://gerarddevalence.substack.com/
