Barclays' September 18 research report argues the humanoid robot bottleneck lies in intelligence rather than hardware, with computing power demand concentrated at the edge rather than in data centers, and the first wave of Physical AI disruption likely not humanoid. Barclays believes large-scale deployment of general-purpose humanoid robots may not occur until around 2035, not 2030, because many demonstrated capabilities still rely on pre-programming, remote control or narrow-domain automation, leaving a significant gap from real general-purpose autonomy. The report cites a "chicken-and-egg" dilemma between scale, cost and capability, similar to autonomous driving's more than ten-year commercialization cycle, and says earlier investment opportunities lie in computing power, data and model layers. On computing power, data centers mainly serve simulation and synthetic data generation and foundation model training and post-training, while reasoning must be completed on dedicated edge processors inside the robot due to strict constraints on latency, power consumption, reliability and security.