Welcome to Autopilot's Industry Breakdown Series. Here, we explore and analyze how AI is reshaping different sectors of the economy. If you’re passionate about understanding the forces driving the future, you’re in the right place. In this review, led by Autopilot Research Fellow Etienne Segal, we delve into the rapidly emerging humanoid robotics industry.
To download the full PDF version of this report, click here.
I. Background
What comes to mind when you think of embodied AI? For centuries, robotics has occupied the realm of science fiction and boundless human imagination. Visionaries like Leonardo da Vinci and Ismail al-Jazari crafted early concepts of mechanical beings, while Jacques de Vaucanson introduced some of the first automatons. In the past century, figures like Isaac Asimov with “I, Robot” and James Cameron’s “The Terminator” shaped our vision of what robots might become in the distant future.
Today, however, that distant future is drawing nearer, driven by rapid advancements in artificial intelligence, falling manufacturing costs, and labor shortages across various industries. These factors have not only accelerated the development of humanoid robots, but also brought us to the brink of a robotics revolution. We will explore the implications of this revolution - how it will shape businesses, the economy, and even household life.
Figure 02 at a BMW factory in Spartanburg SC, August 2024
I believe Optimus will be more valuable than everything else combined at Tesla, because a sentient humanoid robot able to navigate reality and perform tasks on request has no meaningful limit to the size of the economy.
Bipedal robots could radically transform every sector—from housekeeping to eldercare—and free humans from the most undesirable jobs. In 25 years, there could be a billion robots performing a wide range of tasks, ultimately reshaping GDP, productivity, and human happiness.
Humanoid robotics is still in its infancy, but the potential is undeniable. We have spotlighted some of the most promising startups and established companies in the space. These innovators vary widely in their progress, team size, and funding levels, but they all share one goal: bringing humanoids to life.
While it’s tough to predict the industry’s future leaders, industrial applications are clearly leading the charge. Humanoids are poised to make the biggest impact in sectors like manufacturing, logistics, and waste management - industries where labor shortages and dangerous tasks create the perfect opportunity for robots to step in and improve safety and efficiency.
Investment Landscape
Source: Harrison Schell.
The sector has seen billions of dollars invested into robotic developers, infrastructure providers and research labs over the past few years, with venture funds, corporate VCs and angel investors all aiming to take a slice of the pie.
Source: Autopilot Analysis of PitchBook data.
Top humanoid investors by deal count, 2019 - 2024 Source: Autopilot Analysis of Pitchbook Data.
Milestones
II. The Need
As labor markets shift and high-risk industries grapple with safety challenges, the need for robotics and automation is becoming more urgent than ever. The global workforce is shrinking, industries are struggling to fill critical roles, and dangerous jobs continue to put human lives at risk. Humanoids present an interesting solution, offering to bridge labor gaps and enhance safety in hazardous environments.
A. Addressing the Labor Shortage
The labor shortage we’re seeing today isn’t just a short-term issue. According to McKinsey, this shortage reflects long-term demographic trends - fewer young people entering the workforce and an aging population retiring in droves. In the US, industries like manufacturing are feeling this shift the hardest. Deloitte projects that by 2030, the US manufacturing sector will face a shortfall of 2.1 million workers.
B. Enhancing Safety in High-Risk Jobs
Industries like mining, waste management, and nuclear reactor maintenance face significant risks, with high fatality rates and long-term health impacts for workers. These sectors have long been a subject of ethical debate, as the benefits of well-paid jobs are weighed against the inherent risks.
C. Expanding Scientific Exploration
From the extreme depths of the ocean to the frigid poles and volatile volcanic zones, many parts of our planet have remained unexplored. The ocean, for example, covers 70% of Earth’s surface, however we have only explored 5% of it. As we push the boundaries of space exploration, robots will play a key role in missions to the Moon, Mars, and beyond. In the future, we may send robots on long-distance space missions without the need to bring them back to Earth, opening up new possibilities for scientific discovery.
To Humanoid or Not to Humanoid?
The debate around designing robots to resemble humans has been a hot topic in the robotics world. Advocates argue that since our environment, from door handles to stairs, was built for human use, it makes sense for robots to have human-like features. Humanoid robots, they say, could seamlessly perform tasks by fitting naturally into spaces designed for people.
However, critics respond that not every task requires a robot with arms, legs, and a human-like face. Purpose-built machines could be more efficient for certain jobs, like factory automation or specialized medical procedures, where human-like design isn’t necessary and even complicates the given task.
Ultimately, the debate boils down to utility versus versatility. Should we focus on creating many specialized robots for specific tasks? Or should we develop a few humanoid robots capable of adapting to various functions? This decision will shape the future of robotics, influencing cost, scalability, and the overall social impact.
If we want robots to operate in our environments as seamlessly as possible, perhaps those robots should be modeled after people.
Every part, from precision hardware to AI-driven control systems, is crucial in making humanoid robots functional and adaptable. Goldman Sachs Equity Researchestimated that the average cost of building a humanoid robot in 2023 was around $150,000. However, with rapid advancements in technology and increased production, costs have dropped significantly. Today, building a humanoid robot ranges from $30,000 to $120,000. We predict that manufacturing costs will continue to decrease in the coming years, making these machines more accessible for deployment across various industries.
This guide offers a snapshot of the critical components that go into humanoid robots. From advanced motors that replicate human muscle movements to AI systems that enable robots to navigate and respond to their surroundings, these elements work in harmony to create versatile, human-like machines. The table below highlights key categories, suppliers, and their contributions to the overall cost. Understanding the synergy between hardware and software helps shed light on the engineering marvels that are reshaping industries with humanoid robotics.
Humanoid Tech Stack Source: Autopilot Analysis.
As technology continues to evolve, so too will these components, paving the way for a future where humanoid robots play an even greater role across multiple sectors.
Foundation Models and Infrastructure for Robots
Building on the essential components that power humanoid robots, recent advancements in artificial intelligence and infrastructure are redefining how these machines learn and operate. Foundation models in AI are at the forefront, allowing humanoid robots to navigate, adapt, and interact with their environments more intelligently. These models enhance the capabilities of robots, enabling them to tackle complex tasks and unstructured settings with increased autonomy and precision. This shift is opening new possibilities for humanoid robots across industries, as AI-driven systems become more capable of operating in dynamic, real-world environments.
NVIDIA’s Project GR00T is a pioneering initiative aimed at developing general-purpose foundation models for humanoid robots. By utilizing multimodal learning and self-observation, robots can perform tasks based on instructions and past interactions, allowing them to adapt and react in real-time. Powered by NVIDIA’s three-computer robotics stack, Project GR00T is designed to integrate AI-driven reasoning with fast, reactive motion, marking a significant leap in humanoid robot functionality.
Similarly, Google’s RT-2 Models bring vision-language-action integration, enabling robots to understand abstract concepts and apply this knowledge to real-world tasks. By transferring knowledge from the web into robotic actions, these models empower robots to navigate new environments without explicit programming for every scenario.
In the industrial space, Field AI has introduced Field Foundation Models (FFMs), which are purpose-built for unstructured environments. With applications across sectors like construction, oil & gas, and agriculture, FFMs enable robots to autonomously handle unpredictable conditions, making them ideal for operations in the field.
Skild AI, meanwhile, is advancing embodied intelligence through the development of the Skild Brain, a robotics foundation model built to support various robot embodiments, from quadrupeds to mobile manipulators. With a focus on real-world learning, Skild’s approach ensures that robots can react to dynamic environments, paving the way for broader industrial automation.
These innovations in AI infrastructure and foundation models are crucial to the future of humanoid robots, as they continue to evolve beyond their physical hardware to become more intelligent, adaptable, and versatile in the modern world.
IV. The Market
Main Stakeholders
Partnerships
Humanoid robots are gradually being integrated across various industries. In warehouse automation, Amazon and GXO Logistics have deployed Agility Robotics’ Digit, using it to move and sort packages, automating tasks traditionally performed by human workers. Similarly, BMW and Mercedes-Benz have incorporated humanoid robots into their manufacturing operations, aiming to reduce the need for human involvement in repetitive or hazardous roles.
Outside industrial settings, Norway-based 1X has begun piloting humanoid robots in select households, exploring use cases such as home assistance and caregiving. These pilots showcase the potential for humanoid robots to expand into consumer-facing applications, suggesting that they may soon play a role in everyday life beyond industrial applications.
Pricing and Demand Drivers
Currently, pricing for humanoid robots varies widely, with many developers withholding specific figures. Based on available information, prices range from $16,000 to over $150,000, depending on the robot’s complexity, capabilities, and target use case. For example, Tesla’s Optimus is projected to cost between $20,000 and $30,000, while Unitree’s G1 entry-level humanoid is priced at $16,000. On the higher end, Boston Dynamics’ Atlas is estimated to exceed $100,000.
Leasing options, such as labor-as-a-service models, are also emerging. Companies like Agility Robotics and Sanctuary AI offer robots for hourly rates as low as $30, making it an attractive option for enterprises facing labor shortages without the upfront capital investment of purchasing robots.
The demand for humanoid robots is primarily driven by their ability to outperform human labor in speed, precision, and consistency, coupled with reduced maintenance costs. As production scales and technology improves, prices are expected to decrease, making humanoid robots more accessible across various sectors. With advancements in AI and robotics, these machines will become even more capable, driving greater demand for their deployment in increasingly complex tasks.
Labor Market Equation
While the consumer market for humanoid robotics holds great promise, we believe it remains years away. In the near term, industrial applications present the most immediate and impactful opportunity, especially in sectors dealing with labor shortages and high-risk environments. For this reason, we have excluded the consumer segment from our market size calculations and focused on industries where humanoid robots could have the greatest near-term impact.
Our needs assessment has identified several key industries poised to adopt humanoid robots within the next decade. These include manufacturing, warehousing, retail, waste management, and mining - sectors that heavily rely on non-supervisory, repetitive, and physically demanding labor. By using data from the US Bureau of Labor Statistics (BLS), we have estimated the size of the labor markets in these industries, focusing on roles that could be realistically augmented by humanoids
Source: Autopilot Analysis based on BLS data, August 2024.
Relevant Industries at Risk of Humanoid Robot Automation
Retail Retail boasts the largest non-supervisory workforce, with over 13 million employees. This makes it a prime candidate for robotic assistance, where humanoids could assist with stocking shelves, managing inventory, and engaging in CS.
Manufacturing Manufacturing is close behind, employing 9 million workers. With its heavy reliance on repetitive tasks, humanoid robots are perfectly positioned to augment the workforce, boosting efficiency and reducing manual strain.
Warehousing Warehousing is another sector ripe for transformation, where robots can take on physically demanding tasks like lifting and moving goods, as the demand for faster fulfillment surges.
Mining/Waste Management Mining and waste management, with their high-risk environments, stand to benefit from robots stepping in, ensuring human safety in hazardous roles.
Modeling Assumptions
1. Rate of Substitution: By 2035, we assume that over 20% of non-supervisory, repetitive roles in the selected industries will be replaced by humanoid robots. 2. Efficiency Ratio: Initially, humanoid robots may not match human efficiency. We assume a starting efficiency ratio of 0.1 robot hours to human hours in 2024, gradually improving to 1.27 by 2035. 3. Useful Life: We assume a 2-year useful life for humanoid robots before major maintenance or replacement is required. 4. Selling Price: The average selling price of humanoids is expected to drop from $70k in 2025 to $20k by 2035 as manufacturing costs fall and production scales. 5. Workload: We assume robots will be able to work longer hours than humans, starting at 51.5 hours per week in 2025 and increasing to 69 hours per week by 2035, contributing to overall efficiency gains.
Autopilot Analysis
As robot efficiency improves and costs decrease, humanoid robots are expected to replace a significant portion of the labor force in repetitive, non-supervisory roles. This follows Wright’s Law, which states that for every cumulative doubling of units produced, costs will fall by a consistent percentage due to scale efficiencies.
A classic example is Ford’s Model T - as production scaled, costs dropped, enabling wider adoption. Similarly, the manufacturing cost of humanoid robots is projected to decrease as cumulative production increases, making robots more affordable over time. The payback period for robots, which is almost 18 years in 2024, is expected to shrink to less than 1 year by 2035. The broader adoption of humanoids will allow industries reliant on physical labor to benefit from automation on a large scale, mirroring the impact of mass-produced automobiles on society.
Market Size
5.95m (robots deployed in 2035) * $20.4k (robot selling cost in 2035) = $121 billion
Based on our projections, the humanoid robotics market in the US is expected to reach a total market size of approximately $121 billion by 2035, driven by the increasing number of robot workers in industrial sectors. Starting from an estimated market size of $81MM in 2025, the market is set to grow at a CAGR of approximately 108% over the 10-year period from 2025 to 2035.
Challenges
V. The Innovators
As humanoid robots move from the realm of science fiction to real-world applications, a new generation of innovative companies is leading the charge. These developers are harnessing the power of AI, advanced robotics, and human-centered design to solve some of the world’s most pressing labor challenges, from addressing workforce shortages to automating dangerous or repetitive tasks.
We spotlight the Top 10 companies at the forefront of the humanoid robotics revolution. These pioneers are not only pushing technological boundaries but are also shaping the future of industries such as logistics, manufacturing, retail, and beyond.
To view the full company list in gallery or spreadsheet format, access our Airtable file here.
1X Technologies, a Norwegian robotics company backed by OpenAI and EQT Ventures, continues to make waves in the humanoid robotics space. Recently, the company revealed Neo, a humanoid robot designed for home use, capable of assisting with everyday tasks and supporting people with limited mobility. In a teaser video, Neo is shown in a casual interaction, emphasizing 1X’s vision for close, human-friendly robots. Additionally, 1X has secured one of the largest humanoid robot contracts to date, providing 140 service robots to Everon, a US-based commercial security and safety provider, marking a major step in scaling humanoid robots for real-world applications.
Leadership: Bernt Børnich - CEO and Co-Founder Stein Maurice - COO Nguyen Phuong - CSO and Co-Founder ørgen Sundell - Machinist and Co-Founder Pål Løken - Chairman and Co-Founder Christian Rokseth - Partner, Sprettert Erik Ålgård - Co-Founder, Sprettert Kristin Åbyholm - Board Member
Partners and Customers: Everon, NVIDIA
Select Investors: OpenAI, EQT Ventures, Tiger Global, Samsung NEXT
Agility Robotics, founded in 2015 as a spinoff from Oregon State University, is a leader in humanoid robotics. Its flagship product, Digit, is built for warehouse automation and repetitive tasks. GXO Logistics recently deployed Digit in a multi-year deal, marking the first commercial humanoid deployment under a robotics-as-a-service model, with a pricing rate of $30 per hour, offering a return on investment within two years. Agility Robotics plans to produce 10,000 robots annually at its Salem, Oregon facility, addressing labor shortages. The company is led by CEO Peggy Johnson, an ex-Microsoft technology executive.
Leadership: Jonathan Hurst Ph.D - CRO and Co-Founder Peggy Johnson - CEO Pras Velagapudi Ph.D - CTO Aindrea Campbell Ph.D - COO Melonee Wise - CPO Renuka Ayer - CFO Bruce Leak - Co-Founder, Playground Global Matthew Ocko - Co-Founder, DCVC
Partners and Customers: Amazon, GXO Logistics, Ford Motors
Select Investors: Amazon, DCVC, Sony, DARPA, Playground Global
Apptronik, a 2016 spinout from the University of Texas at Austin, is known for developing versatile humanoid robots. Its flagship robot, Apollo, is designed to collaborate with humans in logistics, manufacturing, and patient care. Recently, Mercedes-Benz and GXO Logistics began piloting Apollo in their operations to automate labor-intensive tasks. Built with modular components and NASA expertise, Apollo features linear actuators mimicking human muscle mechanics, swappable batteries, and a force-control architecture for safe human interaction. With backing from NASA’s SBIR program, Apptronik aims to scale humanoid robotics for both Earth-based applications and potential future space missions.
Leadership: Jeffrey Cardenas - CEO and Co-Founder Nicholas Paine Ph.D - CTO and Co-Founder Kay Sheils - CFO Barry Phillips - CCO Luis Sentis Ph.D - Advisor and Co-Founder Stefan Nusser Ph.D - Advisor Gordon Daugherty - Co-Founder, Capital Factory
Partners and Customers: NASA, Mercedes-Benz, GXO Logistics
Select Investors: Plug&Play, Scrum VC, US DoD, National Science Foundation
Boston Dynamics, founded in 1992 as a spin-off from MIT, is renowned for its highly dynamic robots, including Spot, BigDog, and the humanoid Atlas. Acquired by Hyundai in 2021, the company has revolutionized robotics with legged machines capable of performing complex tasks like running, leaping, and even dancing. While its robots have gained viral attention for their agility, Boston Dynamics is now focused on advancing their cognitive abilities through the Boston Dynamics AI Institute. The company aims to commercialize these robots for industries such as logistics and manufacturing. Companies like Otto Group, BMW, Chevron, and Novo Nordisk use Spot, for tasks such as inspection, maintenance, and safety.
Leadership: Marc Raibert Ph.D - Chairman and Founder Robert Playter - CEO Amanda McMaster - CFO Aaron Saunders - CTO Chad Wright - CIO Jason Fiorillo - CLO
Partners and Customers: NASA, Mercedes-Benz, GXO Logistics
Select Investors: Plug&Play, Scrum VC, US DoD, National Science Foundation
Figure AI, a California-based robotics startup, is at the forefront of humanoid robotics with its latest model, Figure 02. Backed by Microsoft, OpenAI, and Nvidia, the company recently raised $675 million in a Series B, valuing it at $2.6 billion. Figure 02 was successfully tested at BMW Group’s Spartanburg plant, performing complex tasks autonomously. The robot features advanced dexterity, sensors, and human-equivalent strength. Figure’s team includes veterans from Boston Dynamics, Google DeepMind, and Tesla, driving innovation for industrial applications. Commercial rollout is planned for 2025, addressing labor shortages and hazardous jobs.
Leadership: Brett Adcock - CEO and Founder
Partners and Customers: BMW
Select Investors: NVIDIA, OpenAI, Jeff Bezos, Microsoft, Intel Capital, LG
Fourier, a Shanghai-based robotics company, recently launched its latest humanoid robot, GR-2, as part of its GRx series. The 175 cm tall GR-2 features 53 degrees of freedom, a detachable battery with a two-hour runtime, and enhanced dexterity with hands offering 12 degrees of freedom. With roots in healthcare robotics and exoskeleton development, Fourier is now focusing on AI-enabled humanoid robots for industrial applications. The company plans to mass-produce its robots by the end of 2023, aiming to deliver thousands of units in 2024. Fourier collaborates with major AI firms to advance the “brain” of its bipedal robots, including NVIDIA’s Project GR00T.
Leadership: Jie Gu - CEO and Co-Founder Zen Koh - CSO and Co-Founder Joseph Chang - MD, Prosperity7 Ventures Suyang Zhang - MD, IDG Capital Yeshun Dong - Investor
Partners and Customers: SensoRehab, ETH Zurich
Select Investors: SoftBank, HongShan (ex Sequoia China), Prosperity7
Neura Robotics, based in Germany, is at the forefront of cognitive robotics innovation. The company’s product line includes collaborative robots, mobile platforms, and its latest humanoid, 4NE-1, designed to interact with humans using AI-powered voice and gesture recognition. Neura has partnered with Nvidia to integrate Jetson Thor for enhanced AI capabilities. The company recently announced a shift in production from China to Germany, emphasizing its commitment to “Made in Germany” quality. With significant advancements in AI and sensor technology, Neura Robotics aims to revolutionize industries from manufacturing to healthcare with cutting-edge robotic solutions.
Leadership: David Reger - CEO and Co-Founder Jens Fabrowsky - COO Bernd Heinrichs - CGO Milad Malekzadeh - Head of AI and Co-Founder Fabian Gruner - Principal, HV Capital Herbert Mangesius - GP, Vsquared Stefan Fritz - Partner, Primepulse
Partners and Customers: NVIDIA, Omron, Jetson Thor
Sanctuary AI, a Vancouver-based robotics company, is developing Phoenix, a general-purpose humanoid robot designed to replicate human-like intelligence. Founded in 2018, the company has raised over $140 million, with backing from BDC Capital, InBC Investment Corp, and Magna International. Sanctuary focuses on solving labor shortages by enabling robots to perform complex tasks in various industries. The company claims to have tested the robots across 400 tasks in 15 different industries, with partners such as Magna and Mark’s retail store. In April 2024, the seventh generation of Phoenix successfully completed 110 tasks in a retail pilot.
Leadership: Geordie Rose Ph.D - CEO and Co-Founder Philip Smith - CFO Paula Gil - COO Olivia Norton - CTO and Co-Founder Ajay Agarwal - Chief Economist and Co-Founder Martin Reed - Partner, Evok Innovations
Partners and Customers: Magna, Ansys, Microsoft, Mark’s Retail
Select Investors: Accenture, Government of Canada, Verizon, DCVC
Announced in 2021, Optimus has been a central focus for Tesla, with Elon Musk predicting it could help drive the company to a $25 trillion valuation and far surpass all its other product lines in terms of value. Equipped with 11 degrees of freedom and tactile sensors, Optimus is built to handle tasks autonomously, such as battery handling in Tesla’s factories. While Tesla plans for low production by 2025 and mass production by 2026, experts view Musk’s timeline as ambitious. Optimus aims to address labor shortages and transform the robotics market, along with the wider economy.
Leadership: Elon Musk - CEO and Co-Founder Vaibhav Taneja - CFO Tom Zhu - SVP Jeffrey Straubel - Co-Founder Kimbal Musk - Board Member
Partners and Customers: Tesla
Select Investors: Fidelity, T. Rowe, Tencent, PIF, Vanguard, BlackRock
Unitree, a Chinese company best known for its advanced quadruped robots, is expanding into humanoid robotics with the development of its H1 and G1 models. The H1 humanoid, standing 71 inches tall and weighing 100 lbs, features 5 degrees of freedom in its legs and 4 in its arms. The more advanced G1 will include 23 degrees of freedom, 3-fingered hands, and a LiDAR-based vision system powered by Intel RealSense cameras. The G1 is designed to handle complex tasks like soldering, delicate object handling, and even flipping food in a frying pan. Unitree has priced the G1 at around $16,000, making it one of the most affordable humanoid robots on the market, ideal for robotics research.
Leadership: Xingxing Wang - CEO and Founder
Partners and Customers: DHL, Accenture, Samsung
Select Investors: Meituan, CITIC, Shenzhen Capital Group, HongShan
Thank you for reading!
Stay tuned for our upcoming pieces in our Industry Breakdown Series.
