2 1 Shi Ji Jing Ji Bao Dao·2025-10-16 09:48Group 1 - The core viewpoint is that in the era of physical AI, large models serve as a new operating system, with data becoming the primary fuel for innovation [2] - The evolution of technology over the past thirty years has highlighted the importance of operating systems, computing power, and data, particularly with the rise of AI and large models [2] - The new generation of AI model companies recognizes that data, operating systems, and computing power all create value, with data often relying on models and operating systems [2] Group 2 - The choice for in-house development is driven by the coupling of the new manufacturing and digital worlds, leading to two forms: full self-research and complete division of labor [3] - Major tech companies, including the "seven giants" in the US and China's BAT and ByteDance, opt for self-research to maximize value [3] - Xiaopeng Motors is testing a new generation of VLA autonomous driving large models and has introduced a new AI system called VLT (Vision-Language-Task/Thinking) for its latest robot [3] Group 3 - The VLT system integrates visual and language inputs to generate task systems, with the potential to evolve into cognitive capabilities [4] - The company is developing a combination of over five operating systems to enhance the capabilities of autonomous robots, transitioning from remote-controlled to autonomous systems [4] - Looking ahead 5 to 10 years, the vision includes autonomous vehicles in cities, flying vehicles between cities, and robots in residential and industrial settings, emphasizing the importance of management and service [4]