According to CoinMeta, a research report by Huatai Securities indicates that in the field of robotics, major companies are focusing on developing embodied brains, and the task planning and generalization capabilities of universal models are rapidly improving. Model capabilities remain the core bottleneck of the industry, but further breakthroughs increasingly depend on real-machine data. To implement these models, issues such as latency and cost still need to be addressed. Data collection and perception, ontology operation control, and high-precision transmission links are expected to benefit first. In the field of intelligent driving, the Ministry of Industry and Information Technology plans to stipulate that autonomous driving system indicator lights can only be optional for vehicles of L3 and above. The boundaries of responsibility for L3 are gradually becoming clearer, and the pace of L3 compliance between 2027 and 2028 will become a key variable in the competition among automakers in intelligent driving. In terms of investment, focus should be placed on companies in the industrial chain that have a customer base and mass production capabilities.