An investor friend sent me a Morgan Stanley report published in April titled Humanoids: A $5 Trillion Market. The report validates my hypothesis that embodied AI (i.e. humanoids) is the next frontier beyond foundational LLMs.

The opening of the report is certainly attention-catching:

Morgan Stanley is making a strategic commitment to telling the story of embodied AI – leveraging our platform and relationships to help our clients identify the next crop of multi-generational compounders transforming industries and creating new markets we believe can exceed the size of today’s global GDP.

History books will be written about this time and the next 5 or 10 years. The implications across markets and geopolitics are likely to be disruptive and far reaching. The Morgan Stanley research team is here for you as we navigate these consequential times. We are grateful for your partnership and thankful for your business. Stay human.

While Nvidia’s Jensen Huang and Elon Musk remain the biggest celebrity promoters of the humanoid technology in the US, the Morgan Stanley report is much more granular and filled with data tables and regression analysis, therefore more insightful to tech followers.

The main takeaways are –

• Humanoid robots could reach 1 billion units by 2050, with 90% used for industrial and commercial purposes
• The humanoids market could surpass $5 trillion, including sales from supply chains, repair, maintenance and support
• Adoption of humanoids is likely to accelerate in the mid to late 2030s with improved technology, and greater regulatory and societal support. It’s likely to achieve mass market penetration in 2040s and 2050s.
• China is leading the race with strong technological prowess, investment, and government support and is the most likely long term winner in the humanoid race

The reasons Morgan Stanley gave for its belief on China’s likely global leadership are –

• Critical supply chain inputs, especially rare earth. According to Morgan Stanley, China can “dial the output of the Western manufacturing complex” with its leverage on rare earth metals and secure key advantages to manufacturing robots as well. Rare earth is a critical input to robot production as Elon Musk publicly acknowledged China’s rare earth export control directly affected the production of Tesla’s Optimus humanoid

According to the bank’s estimates, the rare earth bottleneck is further compounded by the fact that the west could take up to 20 years to build the capabilities to match China’s processing and refining capacity today. In an earlier article, I discussed the importance of rare earth and how China has built an almost unassailable moat around the industry https://huabinoliver.substack.com/p/rare-earth-and-reindustrialization

• Manufacturing capacity and ability to scale. Elon Musk recently posted that manned aerial fighters are irrelevant in future wars since they will be fought with drones. In the next breath, Musk pointed out China makes more drones in a day than the United States makes in a year. While this example is specific to defense production, it’s generally true with all other high tech productions.

Morgan Stanley pointed out the earliest use cases for humanoids will be industrial and commercial. As the largest industrial nation with more manufacturing value added than the combined total of the next four biggest manufacturers (US, Germany, Japan and India), China will lead the adoption of humanoid robots globally.

In 2024, China installed 7 times more industrial robots than the US and over half of total global installed capacity. By 2050, Morgan Stanley estimates China is likely to have 300 million humanoid units, compared with 77 million units for the US.

• Cost advantages. The complexity of humanoids, which require sophisticated robotic software models and tight integration with hardware, makes them an expensive product.

Morgan Stanley Research estimates that the cost of one humanoid was around $200,000 in 2024 in the west. As the technology advances and production volumes increase, prices are likely to fall to about $150,000 by 2028 and $50,000 by 2050. In countries that could tap into Chinese supply chain, prices could fall to as low as $15,000.

As of today, the unit cost of humanoids from Unitree and UBTech, two leading Chinese humanoid makers, is roughly one third of similar products from Tesla or Boston Dynamics.

My personal view is the mature stage Chinese to US price ratio would settle between 1:3 and 1:4. EVs, solar panels, wind turbines and MRI machines are useful proxies, all with similar price ratios.

• Non rare earth supply chain advantage. According to the Morgan Stanley report, China has a self-sufficient industrial supply chain for humanoid product, while there are few US-based alternatives for many humanoid components, such as screws, reducers, actuators, motors and batteries. Nearly every robot developer in the world today requires critical components sourced from China and other parts of Asia.
• Government support. Morgan Stanley reports “every major Chinese city and province has its own fund aimed at embodied AI/robotics”. This support means that Chinese companies are well funded, constantly interlocked in a competition where creative ideas from some lead to the destruction of others. The internal competition is “an underappreciated driver of the rapid pace of AI-robot development in China,” according to Morgan Stanley. Its view reflect the same that I expressed in an article on the Chinese recipe for industrial competitiveness https://huabinoliver.substack.com/p/the-secret-sauce-of-chinese-industrial

Such top-down government support plus marketplace-based competition for critical industries is the Chinese playbook.

• Demographics, public interest and vocational training. Morgan Stanley states “China’s well-known demographic challenges provide a natural incentive to develop technologies in the domain of physical AI”. The country is also generating public interest in robotics through conducting events such as “marathons, boxing competitions, and dance performances.”

According to Morgan Stanley’s research, one of the biggest areas where China leads the US, particularly when it comes to manufacturing robots, is vocational education. The bank points out that China had “5 million students enrolled across over 11,000 vocational schools” in 2023, while data from the “National Student Clearinghouse Research Center recently estimated that there are 900,000 students enrolled in vocational-focused schools in the US.

On the human capital level, beyond vocational training, China graduates 5 times STEM college students a year than the US, 8 times STEM PhDs.

• “Long Game”. The final Chinese advantage, as per the report, is the tendency to play the “Long Game”. Using the example of the Chinese board game Go, Morgan Stanley believes China’s approach is built on “patience and combative coexistence” to foster national champions in a fiercely competitive marketplace. The bank comments that Chinese strategic thinking is based on principles dating back to the fifth century BC, while the US “is a much younger country” where social mobility “can skew companies and investors towards short-term thinking, prioritizing immediate results (near-term growth, margin expansion, buybacks, etc.) over long-term strategic planning.”

In conclusion, “it is becoming apparent that national support for ‘embodied AI’ may be far greater in China than in any other nation, driving continued innovation and capital formation,” says the Morgan Stanley report. “In our opinion, China’s lead in AI-robotics may widen with rivals, including the US”.

“While it is too soon to declare a final champion in the race for agentic humanoid robot supremacy, the US will need to make significant changes in manufacturing capability, education and national policies to remain competitive in this area,” the report says.