Overview

This report evaluates the degree of overlap between intelligent vehicles and humanoid robots in core products and technology stacks across three dimensions: perception layer, decision-making layer, and execution layer. It further analyzes the key differences between the two industries in functional safety standards, BOM cost structures, and mass production timelines. Finally, it outlines the trends of technological convergence and industrial divergence over short-, medium-, and long-term horizons.

What the Full Report Will Cover

This report provides a systematic analysis of technological convergence and industrial divergence between intelligent vehicles and humanoid robots along the perception–decision–execution technology stack. It assesses the degree of hardware and software reuse at each layer—sensors, perception algorithms, VLA/world-model decision architectures, AI chips, simulation platforms, and actuation systems—with quantitative benchmarks on sensor counts, degrees of freedom, and BOM cost structures. The report then examines the strategic entry of nearly 20 Chinese automakers into humanoid robotics, classifying their approaches into four distinct playbooks: aggressive full-stack reuse, scenario-driven pragmatism, supply-chain leverage, and early-stage exploration. Finally, it identifies three structural fault lines—functional safety standards, BOM cost composition, and mass-production timelines—that ensure long-term industrial separation despite shared technology foundations, and outlines convergence and divergence trajectories across short- (1–3 year), medium- (3–5 year), and long-term (5–10 year) horizons.

Key Questions the Report Answers

• How much of the intelligent-vehicle technology can be repurposed for humanoid robots, and where do the limits lie?
• Where does sensor reuse highest and where is it essentially zero, and what does that imply for component suppliers' dual-track growth?
• Why do autonomous driving and robotic perception follow such different paradigm paths, and when might they converge?
• How does the order-of-magnitude gap in degrees of freedom reshape the economics of VLA model deployment, chip design, and simulation infrastructure across the two industries?
• Why will automotive-tier functional-safety certification, a powertrain-dominated BOM, and ten-million-unit annual scale prevent the two industries from simply merging into a single "embodied intelligence" category, and what are the investment implications at each phase of the convergence-divergence dynamic?