远景科技集团董事长张雷：美国搞不定的能源大模型，我们三年内做大做强

Core Concept - The concept of "Physical AI" is introduced as a new paradigm that combines AI with physical laws and knowledge graphs, aiming to eliminate the "hallucinations" of traditional language models and enable reliable AI applications in the physical world [2][3]. Group 1: Development of Physical AI - The future development of Physical AI is seen as a significant direction, particularly in the context of energy systems [2]. - The integration of data intelligence with physical laws like energy conservation and aerodynamics is expected to enhance AI's reliability in real-world applications [2]. Group 2: Energy Model and AI Applications - The energy model is considered crucial for reconstructing energy systems, providing a rich application scenario for Physical AI [4]. - AI's ability to process vast amounts of data in milliseconds can help optimize decision-making in complex energy systems, addressing human anxieties related to energy management [4][5]. Group 3: Competitive Landscape - The U.S. is viewed as lacking the necessary industrial scenarios and complex energy systems to support the development of Physical AI and energy models, giving China a potential advantage in this field [3][5]. - Companies that only specialize in single areas like wind or solar energy may struggle to develop comprehensive energy models due to a lack of holistic understanding and data [5]. Group 4: Addressing Industry Challenges - The energy sector is currently facing issues of overcapacity and price wars, particularly in solar and wind energy, which have led to significant financial losses for many companies [7]. - Physical AI and energy models are proposed as solutions to end the cycle of homogeneous competition and shift the focus from material assets to intelligent assets [8]. Group 5: Future Outlook - The development of energy models is expected to evolve into a robust system capable of generating significant value within 1-3 years [11]. - The future energy system is envisioned as an ecosystem of intelligent agents rather than just a collection of devices, aimed at better integrating renewable energy sources and providing energy at lower costs [11].