
Bernstein Research Report Analysis: Payback Period for Four Major Humanoid Robot Scenarios Drops Below 2.5 Years, Global Shipments Projected to Reach 1 Million Units by 2031
By Rita
The humanoid robot industry is undergoing a critical turning point. Market focus has shifted from "whether they can walk" to "whether they can make money".
On August 25, Bernstein released a research report quantifying for the first time the economic viability of humanoid robots across four major application scenarios. The report calculates that in urban patrol, warehouse palletizing, factory material handling & inspection, and last-mile delivery across the US and China, the payback period for patrols is already under two years. For other scenarios, once prices drop by 30% to 50%, payback periods are expected to fall below the 2.5-year threshold for accelerated adoption.
Bernstein forecasts that global annual shipments of humanoid robots will reach 1 million units by 2031, with the four key scenarios accounting for approximately 50% of that total.
Patrols Are Already Profitable; Other Scenarios Require Further 30%-50% Price Cuts
Bernstein breaks down the humanoid robot payback period into "initial investment" divided by "annual net savings". Initial investment covers robot procurement costs and system integration fees, while annual net savings derive from labor cost displacement minus operations maintenance and manual supervision expenses.
At current pricing, payback periods across the four scenarios range from 1 to 6.7 years. US urban patrols have the shortest payback period at just one year, while Chinese factory material handling & inspection has the longest at 6.7 years.
US payback periods generally outperform those in China. Labor costs in the US are three to five times higher than in China; although robot selling prices are higher, the displaced labor costs allow initial investments to be recouped much faster.

Urban patrols have already achieved payback within two years in both the US and China. To reach the 2.5-year payback threshold, the other three scenarios—warehouses, factories, and delivery—require robot price reductions of 30% to 50% in the Chinese market and 5% to 40% in the US market.
Bernstein believes pricing will not be a long-term barrier. Experience from the EV and lithium battery industries shows that every doubling of production volume drives cost reductions of 15% to 18%, with steeper cost curves during ultra-low-volume stages. Once humanoid robots enter mass production, cost reductions may outpace expectations. Major manufacturers like Agility have already showcased similar downward cost projections in investor presentations.
Potential Market Across Four Scenarios Exceeds 10 Million Units
Bernstein estimates that the four scenarios could accommodate approximately 10 million humanoid robots in China and 1.2 million in the US. Factory material handling & inspection represents the largest single scenario, with deployable capacities of 6 million in China and 600,000 in the US. Urban patrol accounts for 1.8 million in China, warehouse palletizing 500,000, and last-mile delivery 1.6 million.
Assuming an 8% penetration rate within five years, cumulative installations across the four scenarios will reach approximately 900,000 units by 2031, with annual shipments exceeding 500,000 units, representing roughly 50% of the global total for humanoid robots that year.
Bernstein acknowledges this does not represent the entire humanoid robot market. Longer-term potential markets for scenarios such as factory assembly, commercial services, and home applications are even larger. While proof-of-concept pilots have emerged, these areas still lack the technological maturity and economic viability for large-scale deployment in the near term.
Three Metrics to Identify Winners
Bernstein advises investors to track the competitiveness of humanoid robot companies across three dimensions.
Growth trajectory of actual commercial sales. Proof-of-concept and lab demonstrations cannot form a sustainable business model; verifiable expansion among commercial customers is the core metric for determining whether a company has genuinely entered the commercialization stage. Growing orders and repeat purchases are more convincing than any technical demonstration.
Performance improvement trajectory. The volume of data matters less than the efficiency of the data closed-loop. Companies capable of continuously optimizing reinforcement learning models using real-world deployment data will see significantly faster performance gains than competitors stuck at the lab stage. The key lies in whether a closed-loop system for "training models with data" is established and operating continuously.
Trend toward the center of the industrial ecosystem. The humanoid robot supply chain spans chips, sensors, motors, reducers, AI models, and multiple other components. Robot companies that attract system integrators to build ecosystems around themselves will achieve faster iteration speeds and lower supply chain costs.
Humanoid robots are transitioning from technical demos to business cases. Once payback periods drop below 2.5 years, procurement decision-making logic will shift from "should we try it?" to "why aren't we buying it?" By then, the slope of the shipment growth curve will be steeper than anticipated.

Disclaimer
This article compiles and interprets a third-party brokerage research report (Bernstein, August 25, 2026) by Chaoxiang Research, along with publicly available market information. The ratings, target prices, earnings forecasts, and related judgments cited herein reflect only the views of the brokerage's analysts, represent solely the stance of their respective institutions, do not constitute the views of Chaoxiang Research, and should not be construed as any investment advice.
Markets involve risks, and decisions must be made independently. This article should not serve as the basis for buying or selling any securities.















