2026.02.03 本文字数：1699，阅读时长大约3分钟 作者 |一财评论员 从制造到智造，AI重塑制造业生态正步入实战阶段。 在此过程中，充分吸收各地产业布局的成功经验至关重要。这方面如何充分吸收各地在产业布局方面的 成功经验，将是一个重要的抓手。以新能源车为例，作为燃油车重镇的上海引进特斯拉，以国际化视野 和产业前瞻布局，成功带动长三角打造新能源汽车的国际产业链供应链。这为中国制造迈入中国智造， 为利用人工智能技术对生产制造全链条全周期重塑，提供了宝贵的经验。 必须强调的是，上海等地的成功实践，其核心经验之一在于构建了自由开放的经济环境。中国新能源汽 车产业形成的国际竞争优势，正是在开放市场的基准场景下，汇聚全球技术、标准与智慧的结果。这一 逻辑在物理世界AI时代将变得更加重要。因为未来制造业供给的更多是AI Agent（智能体），这种商品 即服务的智能体，是不断迭代升级的"半成品"，这将给商品安全、技术标准，特别是智能体如何适应不 同属地法律带来全新挑战，使得国际贸易规则和市场准入标准趋于复杂。 这意味着，物理AI时代的市场竞争，将更加依赖于市场的开放性、技术的开源性和制度的包容性。物 理AI时代的 ...

专题：2026年度国际消费电子展(CES) 新浪科技讯 1月6日上午消息，一年一度的CES展会期间，黄仁勋在英伟达新品发布会上表示，一个"物 理AI时代"正在开启。他描绘了一个充满想象力的未来：自动驾驶汽车具备推理能力，机器人能够理解 并思考，AI Agent（智能体）可以处理百万级token的长上下文任务。 他说，像DeepSeek R1这样的开源推理模型，让整个行业意识到：当开放、全球协作真正启动后，AI的 扩散速度会极快。尽管开源模型在整体能力上仍比最前沿模型慢大约半年，但每隔六个月就会追近一 次，而且下载量和使用量已经呈爆发式增长。 新浪科技讯 1月6日上午消息，一年一度的CES展会期间，黄仁勋在英伟达新品发布会上表示，一个"物 理AI时代"正在开启。他描绘了一个充满想象力的未来：自动驾驶汽车具备推理能力，机器人能够理解 并思考，AI Agent（智能体）可以处理百万级token的长上下文任务。 责任编辑：江钰涵 专题：2026年度国际消费电子展(CES) 他说，像DeepSeek R1这样的开源推理模型，让整个行业意识到：当开放、全球协作真正启动后，AI的 扩散速度会极快。尽管开源模型在整体能力上仍 ...

Core Points - The 15th National Games will be held from November 9 to 21 in Guangdong, Hong Kong, and Macau [1] - The torch relay for the event started on November 2, with a total of 200 torchbearers, each city having 50 torchbearers running 50 to 100 meters [1][3] Group 1: Torch Relay Details - The torch relay will conclude with representatives from the four cities returning the flame to Guangzhou, followed by a ceremonial fire fusion event at the Guangdong Olympic Sports Center [3] - A humanoid robot named "Kua Fu" served as the "0th torchbearer," successfully completing a 100-meter torch relay task [5] - The robot was developed by Shenzhen Leju Robotics and underwent enhancements in dynamic motion control, load balancing, and complex environment adaptation [5] Group 2: Technological Innovations - The robot's operation was optimized using a 5G-A network, achieving an uplink speed of 20 Mbps and a latency of 20 ms for stable communication [6] - This technology allowed for real-time transmission of ultra-high-definition footage from the robot's perspective, enabling precise execution of relay commands without the need for accompanying personnel [6] Group 3: Athlete Participation and Community Engagement - Notable athletes, including Quan Hongchan, are participating in the diving events, with hopes of increased local audience attendance [8] - Liu Weiqiang, the opening ceremony's director, expressed excitement about the event and emphasized the close ties among Guangdong, Hong Kong, and Macau [10][12] - He also encouraged athletes from other regions to explore and enjoy the local culture and cuisine [12] Group 4: Personal Reflections from Torchbearers - He Xiaopeng, the final torchbearer in Guangzhou, expressed pride and excitement about participating in the relay, linking the event to technological innovation and the future of AI [14] - He believes that advancements in technology will significantly improve lives in the coming decades [14]

Core Points - The 15th National Games will be held from November 9 to 21 in Guangdong, Hong Kong, and Macau [1] - The torch relay for the event started on November 2 in Shenzhen, featuring 200 torchbearers across four cities [2][3] Group 1: Torch Relay Details - The torch relay involves 200 torchbearers, with each city running 50 torches over distances of 50 to 100 meters [2] - A unique feature of this relay is the participation of "Kua Fu," a humanoid robot, as the "0th torchbearer," marking a first in the use of a 5G-A humanoid robot for such an event [3][4] Group 2: Technological Innovations - "Kua Fu" successfully completed a 100-meter torch relay task with a load of 1.6 kg, showcasing advancements in dynamic motion control and load balancing [4] - The robot's performance was enhanced through collaboration with Harbin Institute of Technology and Beijing General Artificial Intelligence Research Institute, focusing on motion control algorithms [4][5] Group 3: Communication Infrastructure - China Mobile optimized over 20 stations along the relay route using 5G-A technology, achieving an uplink speed of 20 Mbps and a latency of 20 ms for stable communication [5] Group 4: Athlete Participation and Community Engagement - Notable athletes, including Olympic champion Yang Jinghui and diver Quan Hongchan, are participating in the games, encouraging local audiences to support them [6][8] - Liu Weiqiang, the director of the opening ceremony, expressed excitement about the event's significance in strengthening ties among Guangdong, Hong Kong, and Macau [10] Group 5: Perspectives on Innovation - He Xiaopeng, chairman of Xiaopeng Motors, emphasized the importance of technological innovation, likening it to the torch's role in illuminating the future [11]

Core Insights - The recent collective study by the Central Political Bureau on artificial intelligence (AI) highlights its strategic importance as a transformative technology that is reshaping human production and lifestyle [1] - Despite significant advancements in AI capabilities, there remain critical weaknesses in foundational theories and core technologies [2][4] Group 1: Foundation Research - The necessity for breakthroughs in foundational theories, methods, and tools is emphasized as essential for gaining a competitive edge in AI [2] - The concept of "root technology" is introduced, with the Transformer architecture, which underpins popular models like ChatGPT, being a prime example [2] - Current limitations in AI include insufficient advanced differential computing capabilities, lack of scientific reasoning, and issues with interpretability and stability [2][3] Group 2: Key Technology Risks - There exists a "bottleneck" risk in critical core technologies, particularly in AI chips, where domestic GPU performance lags behind global leaders like NVIDIA by approximately three years [4] - The dominance of U.S. institutions in foundational algorithms poses a challenge, as most foundational models originate from American universities and tech companies [4][5] - Data isolation and a lack of high-quality data are significant barriers for AI model training in China [4] Group 3: System Deployment and Recommendations - Urgent action is required to enhance foundational research in AI, with recommendations for increased national support in mathematical and computational theories [6] - The focus of foundational research should be on physical world modeling and the construction of intelligent robotic systems, which are crucial for reshaping industry competition [6][7] - A call for a collaborative effort to build an open-source platform and establish a "Physical AI Open Source Alliance" is made to promote international data standards [7] Group 4: Future Directions - The future of AI breakthroughs is expected to concentrate on three areas that deeply integrate with the physical world [7] - Recommendations include supporting hardware-neutral platforms, establishing national pilot zones for practical applications, and creating replicable models that combine digital twins with the real economy [7]