Humanoid robots: ideal is rich, reality is skinny? Industry predicts 100,000 units sold next year, institutions only see 12,000 units.
Morgan Stanley warns that the market's short-term expectations for humanoid robots are overly optimistic, and true large-scale commercial adoption faces a triple challenge: application scenarios, hardware costs, and software intelligence.
According to TrendChaser Trading Desk, Morgan Stanley stated in a November 12 report that currently there is a huge divergence between industry companies and analysts regarding future sales of humanoid robots. At last week's GGII Humanoid Robot Conference, several companies predicted that China's demand for humanoid robots will reach 100,000 units in 2026. This is in sharp contrast to Morgan Stanley analysts’ forecast: only 12,000 units in 2026, rising to just 114,000 units by 2030.
Behind this nearly tenfold difference in expectations is deep concern among institutions about the current limited working capabilities of robots, their unattractive return on investment (ROI), and a series of product development challenges.
Morgan Stanley analysts emphasize that the path to large-scale deployment of humanoid robots is not easy. The development of this technology will not happen overnight; rather, it will require a long and arduous uphill process. In the short term, compared to the much-hyped “general-purpose humanoid robots,” companies making specialized robots (such as wheeled or quadruped robots) that can create clear value in specific tasks – as well as core component suppliers – may be more worthy of attention.
General-purpose is a distant goal; specialization in specified scenarios is key
Morgan Stanley believes that although the ultimate goal of humanoid robots is to become general-purpose platforms, the reality is that their “brains”—artificial intelligence—still needs time to evolve.
Analysts indicate that, in the short term, applications will be highly focused in specific, structured B2B vertical sectors. The first things to appear will not be all-powerful robotic butlers like in movies, but robots handling particular tasks in commercial services (such as guiding, performing) and certain industrial scenarios.
Compared to industrial manufacturing, which requires high efficiency and precision, commercial service scenarios demand less from robots; here, robots mainly provide “emotional value,” making them easier to deploy in the early stages. Morgan Stanley expects the industry will need 3–5 years to expand from these narrow use cases to broader fields.
Debate over application scenarios: Industrial manufacturing ROI is uncertain, commercial services may be a “safe haven”
There is intense debate within the industry regarding the first large-scale application scenario for humanoid robots.
Although labor shortages spark imagination for industrial applications, the harsh reality is that the current work efficiency of humanoid robots is only 20–30% that of humans.
Such low efficiency, coupled with concerns about integration complexity, thermal management, reliability, and safety, cast significant doubt on their short-term ROI in industrial fields.
In sharp contrast, commercial service areas (such as mall guides, dance performances), which require lower efficiency and precision, are seen as more realistic early breakthroughs.
For investors, this means they need to carefully scrutinize companies claiming they will disrupt manufacturing with humanoid robots in the short term; their commercial feasibility may be far less than anticipated.
Premature “involution” could stifle industry growth
Cost is a key driver for commercialization, and customers expect finished robots to be priced around RMB 100,000–200,000.
However, Morgan Stanley issues a crucial warning: Excessive focus on price and entering a “price war” too early may sacrifice product reliability and performance stability, ultimately harming the health of the entire industry.
Historical experience shows that successful automation technology rollouts are always “value-driven before cost-driven.”
Therefore, at this stage, the priority should be delivering real functional value and justifying early price premiums with differentiated solutions. Only when the technology matures and scale effects kick in will cost reduction become sustainable.
Hardware bottleneck: visible “body” issues
The “body” of humanoid robots also faces numerous challenges.
First, there is a structural disconnect between system integrators and component suppliers. As the industry is still in its early stages and products iterate rapidly, component performance specs remain unclear, standards are fragmented, and customization costs are high.
Second, component quality consistency and yield rates are major obstacles to commercialization.
Morgan Stanley reports that currently, core components such as joints and actuators still rely heavily on manual assembly, resulting in inconsistent performance and overweight designs. Given close interactions with humans, safety standards for humanoid robot components should surpass automotive standards. Resolving these issues and transitioning from manual assembly to fully automated production is crucial.
Software bottleneck: invisible “soul” restraints
If hardware is the body, then models and algorithms are the robot’s “soul”—and this is currently the biggest bottleneck.
To achieve GPT-level general robotic capabilities, trillion-scale multimodal data including real-world force and contact information are needed for training, but such datasets are extremely scarce for now.
Although “Sim2Real” is frequently discussed, its effectiveness is still under debate. Additionally, computing power is a structural bottleneck; future models may require thousands of TOPS, far beyond the cost and power limits of current edge devices. In summary, breakthroughs in software and models are highly uncertain and represent the core risk investors must carefully consider when evaluating related companies.
Risk warning and disclaimerThe market is risky; investment should be cautious. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, views, or conclusions in this article are suitable to their specific circumstances. If you invest based on this, you do so at your own risk.