Humanoids – robots with two arms, two legs and an AI brain – could become more common in many industries. And it might happen sooner than you expect. Goldman Sachs estimates the addressable humanoid market will reach $38bn by 2035.
They could help ease labour shortages, lower operating costs and boost productivity across multiple sectors, potentially having transformative impacts on businesses and investors.
More than 4 million industrial robots already weld, paint and perform other tasks in factories worldwide. And that’s not counting the robots working in logistics.
Amazon employs more than 1 million robots in its warehouses. Most of these robots aren’t humanoids and lack their capabilities. However, as technology progresses, widespread adoption of humanoids is getting closer to reality.
This shift might not happen overnight, and significant hurdles could slow or even stop its progress. But we think it’s just a matter of time before humanoids reach critical mass, benefiting robotics companies and businesses using these machines.
Benefits
Humanoid robots are designed to do more with fewer constraints than many industrial robots currently in use.
Those industrial robots are essentially disembodied arms dedicated to specific tasks, like picking up and sorting items. They remain stationary in one location. Others resemble amped-up vacuum cleaners, but instead of cleaning floors, they carry inventory throughout a warehouse.
These machines perform tough, boring jobs efficiently. But they’re not very versatile. Being “human-shaped” is another advantage for humanoids. After all, AI-powered robots, such as those from Waymo and Skydio, exist today as autonomous cars and drones. In contrast, some humanoids can carry boxes, stack empty containers and even adjust their routines if there are backups.
However, because humanoids are about the same size as humans and move like we do, they can operate more easily in buildings designed for people. No expensive retrofits are required.
Reasons behind the rise
The world population is still growing, but not evenly. According to the United Nations, dozens of countries, including China, Germany, Russia and Japan, have already peaked.
Deploying humanoids to address population challenges could seem compelling to businesses and policymakers.
Also, more companies might reshore their production if higher US tariffs and protectionism become permanent. Right now, a labour shortage is hindering these efforts. However, the mass availability of affordable humanoids could change the calculation for companies considering a move.
Technology is advancing to the point where robotics companies can build humanoids capable of performing a wider range of tasks.
The tech is improving in three specific areas:
Artificial Intelligence (AI): Humanoids are essentially ‘embodied AI’, a way for artificial intelligence to operate in the physical world. AI acts as a robotic brain, allowing humanoids to navigate a crowded factory, learn new tasks and respond to verbal commands. Until AI evolves, though, humanoids might do best in relatively limited and structured jobs.
Actuators and Sensors: Humanoids now have better ‘eyes and hands’ thanks to developments in sensors and actuators. Actuators generate the physical force to move a robot’s limbs or hands. Newer generations possess a wider range of motion and can make more precise movements.
Sensors capture a massive amount of data as a humanoid engages with its environment. Video, LIDAR, accelerometers and pressure sensors collect information about the robot’s speed, position and grip strength, as well as the objects around it. Without these inputs, humanoids can’t function smoothly in the physical world.
Actuators and sensors have room for improvement, especially in precision. Companies are also working on developing strong and lightweight materials for everyday use.
Batteries: Rechargeable battery life is also improving, which could allow humanoids to move around more freely and work longer. Figure recently unveiled its third-generation battery, which can run for five hours at peak performance. Five hours isn’t a full shift. Batteries are getting better, but industrial users may demand more.
Obstacles
Robotics companies aren’t producing humanoids at scale. Figure recently launched its BotQ production facility, which could produce 12,000 robots in 2025. The company still wants to raise its output substantially.
Part of the problem is the supply chain, or the absence of one. Figure said it had to design “almost the entire robot from scratch” because there isn’t a mature supply chain.
According to Morgan Stanley Research, building a humanoid can cost up to $300,000 on the high end, though some firms are quoting lower prices. Unitree, a Chinese robotics company, says its Unitree G1 humanoid starts at $16,000.
If companies can’t mass-produce humanoids, achieving the necessary economies of scale will be harder. This could keep prices higher, further limiting adoption.
Supply chains may eventually grow to meet this need. Given their expertise, we see China and Japan having an advantage in this area.
Elsewhere, humanoid robots are designed to take over jobs that humans do now. That may lead to resistance from consumers, voters and elected leaders, even if the robotics industry ultimately creates more jobs than it destroys.
Labour unions have also voiced concern about more automation. More opposition may spring up. Historically, though, societies usually embrace innovations that are net positives, even if they have drawbacks.
Conclusion
A few mega-cap companies currently dominate the field of artificial intelligence; however, this hasn’t yet occurred in the realm of humanoid robots. Conditions may favour mid-cap firms in the short term, particularly within the industrial sector.
This includes mid-size companies that specialise in the components humanoid robots require, such as analogue semiconductors, motors and power systems.
Companies will also need large and specialised supply chains to manufacture humanoids at scale, potentially creating opportunities for mid-cap firms.
Establishing these systems may take years, and perfecting them could take decades. However, the long-term trend calls for more humanoids, possibly sooner than many expect.
Nalin Yogasundram and Jonathan Bauman are portfolio managers at American Century Investors. The views expressed above should not be taken as investment advice.