The three major controversies surrounding humanoid robots: When will they be deployed on a large scale? Will computing power demands surge? Are humanoid robots necessary?

Barclays believes that large-scale deployment of general-purpose humanoid robots may not occur until around 2035, rather than 2030.

In its latest research report released on September 18, Barclays made three judgments on the investment boom in humanoid robots: the bottleneck lies in intelligence rather than hardware, computing power is concentrated at the edge rather than in data centers, and the first wave of disruption from Physical AI may not be humanoid at all.

In the report, Barclays also provided a detailed analysis of the issues that investors care about most, such as scale, computing power, and form factor.

Controversy 1: Large-scale deployment, 2035 rather than 2030

Barclays wrote in its report that many of the humanoid robot capabilities currently being demonstrated in the industry still rely on pre-programming, remote control, or narrow-domain automation for specific scenarios, and are significantly far from true general autonomy.

The organization believes that the core bottleneck lies not in hardware, but in intelligence—the AI model capabilities required for perception, reasoning, and action are still immature. On the hardware side, there is a "chicken and egg" dilemma between scale, cost, and capability: without scale, there is no low cost; without sufficient intelligence, commercial value cannot be proven; and it is difficult to achieve breakthroughs in all three simultaneously.

This situation is highly similar to the more than ten-year commercialization cycle that autonomous driving has gone through.

Therefore, Barclays believes that earlier investment opportunities lie in the computing power, data, and model layers—the underlying infrastructure needed to unlock the "GPT moment" for humanoid robots. Large-scale economic deployment of hardware, however, requires breakthroughs in the intelligence layer first.

Controversy 2: Computing power demand – Data center growth is limited, the edge is the main battleground

The impact of humanoid robots on computing power needs to be viewed from two dimensions: centralized data center computing power and distributed edge AI computing power.

On the data center side, computing power primarily serves two types of tasks: simulation and synthetic data generation (for training and testing strategies), and the training and post-training of basic models (building intelligence from perception to action). However, inference—the part where robots perceive, decide, and act in real-time in reality—must be performed on dedicated edge processors within the robot due to stringent constraints on latency, power consumption, reliability, and security. This means that the inference workload of humanoid robots will not directly translate into incremental demands for hyperscale data centers like Agentic AI. For data centers, humanoid robots represent a small incremental increase; for edge computing, however, it could be a significant driver.

However, in an industry where AI data center capacity is already strained, even incremental growth is significant. More importantly, computing power demand precedes hardware deployment expansion. Developers need massive computing power for simulation and model training before robots can be deployed on a large scale.

Figure's partnership with emerging cloud service provider Nscale is a prime example. This multi-year agreement involves an initial investment of approximately $3.5 billion, with a target size exceeding $6 billion, and can support up to 100,000 NVIDIA Vera Rubin GPUs. Based on an estimated 3.0-3.3 kilowatts of rack capacity per GPU in a fully configured configuration, this corresponds to approximately 300-330 megawatts of IT capacity—even before humanoid robots are widely deployed, the development of the "brain" alone already represents a considerable incremental demand for data centers.

Controversy 3: "Humanoid" may not be the optimal solution for Physical AI.

Barclays believes that the first wave of disruption from Physical AI will most likely not be humanoid.

Barclays' statement that it's "too early" specifically refers to fully autonomous general-purpose humanoid robots. Prior to this, AI robots designed for specific tasks have already been deployed across multiple industries. Collaborative robots (cobots), autonomous mobile robots (AMRs), AI drones, quadruped robots—these diverse robots are starting from "dirty, dull, and dangerous" jobs and expanding into various scenarios including commerce, industry, and defense.

Amazon currently has over 1 million robots in operation, encompassing mobile drive units, automated mobile robots (AMRs), and AI control systems. These robots are organized by task and form, and none are humanoid. Atoms, founded by Uber founder Travis Kalanick, recently completed a $1.7 billion funding round, focusing on specialized industrial robots and physical AI systems for the mining and transportation sectors.

So why pursue a humanoid form? Barclays' analysis suggests two main reasons:

First, the human world is built around the human body. Stairs, doors, tools, workbenches, and production lines are all designed to human scale, allowing humanoid robots to directly enter existing environments without requiring the reconstruction of infrastructure.

Second, general-purpose humanoid robots offer the possibility of a multi-tasking platform—the same robot can handle materials, operate tools, and inspect equipment. The value lies in a platform that covers multiple scenarios, rather than outperforming dedicated robots in any single task.

However, when the task and environment are clearly defined, specialized solutions are often faster, cheaper, and safer. In a flat warehouse, wheels are more practical than bipedal vehicles; on uneven terrain, quadrupeds are more stable than bipedal vehicles; and in repetitive operations, specialized grippers are more precise than five-fingered hands. Ultimately, the choice of form factor is determined by the specific scenario.

~~~~~~~~~~~~~~~~~~~~~~~

The above content is from Zhuifeng Trading Platform .

For more detailed analysis, including real-time updates and firsthand research, please join the [ Trading Channel Annual Membership ].

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.