Core Viewpoint - The relationship between Artificial Intelligence (AI) and Environmental, Social, and Governance (ESG) is interdependent, with AI's future development relying heavily on ESG principles to ensure sustainable growth and societal benefits [6][15]. Group 1: AI and Energy - AI's advancement is closely linked to energy resources, with a new trend suggesting that future energy needs may be met through space-based solar power, reducing reliance on terrestrial resources [7][16]. - The concept of "the end of energy is green energy" is evolving, indicating a shift towards sustainable energy solutions for AI operations [7][16]. Group 2: Resource Utilization - AI is driving a transformation in resource usage, moving from traditional materials to more sustainable composite materials, influenced by ESG's stringent resource consumption requirements [8][17]. - Innovations in technology are enabling more efficient resource utilization, particularly in industries like electric vehicles, which are transitioning to safer and cheaper materials [8][17]. Group 3: Energy Efficiency - AI technologies are exploring energy-efficient solutions, such as optical computing, which significantly reduces energy consumption compared to traditional electronic signals [8][18]. - Advances in chip technology are also contributing to lower energy usage, aligning with ESG's demands for improved energy efficiency in AI hardware [8][18]. Group 4: Social Responsibility - The development of AI must address social challenges, including aging populations, by adhering to ESG principles that promote social equity and worker rights [9][18]. - Leading AI companies are already taking steps to ensure that their technologies benefit society as a whole, not just capital interests, by exploring solutions like universal basic income for the elderly [9][18]. Group 5: ESG Ratings and Initiatives - The Sina Finance ESG Rating Center provides various services to promote ESG principles among listed companies, enhancing their sustainable development performance [1][10]. - The center aims to establish ESG evaluation standards suitable for China's context, fostering the growth of ESG investment in the asset management industry [10][19].

此外，热力学芯片等新型硬件通过多态单元设计，进一步降低功耗。王忠民强调，"这些创新的背后， 是ESG对能效的要求促使AI硬件变革，让AI能在低耗环境中持续发展。" AI不仅关乎技术竞争，更直接关联社会结构。王忠民表示，随着AI推动寿命延长，老龄化等问题日益 紧迫。"AI在解决这些问题时，必须遵循ESG中的社会责任要求，关注劳动者权益、财富分配与社会可 持续发展。" 他提到，已有许多领先的 AI 公司开始采取行动。例如，AI可以探索如何为老年人提供更好的保障，如 建立全球基础收入系统，为老年人提供足够的生活保障，让他们在享受健康长寿的同时，不会因失去工 作能力而陷入贫困。"这些都是ESG对AI提出的社会关怀要求，确保AI的发展不仅服务于资本，也服务 于每一个普通人，促进社会的和谐与公平。" 在资源维度上，王忠民指出，AI的发展正推动材料使用方式的根本变革。传统产业依赖对单一材料的 高耗能提纯，而ESG对资源消耗的严格要求，促使AI向更可持续的模式转型。"从单一材料到复合材 料，再到化合材料，技术进步让我们更高效地利用资源。" 他举例说，新能源汽车电池正从依赖锂、钴等稀有金属，转向更廉价、安全的复合材料，"这不仅 ...

Core Insights - Extropic is developing a new type of computing chip based on thermodynamic computing, which aims to utilize random thermal fluctuations rather than minimize them as in quantum computing [1][2] - The company plans to launch its first chips in late 2025, with expectations that they will be capable of running complex workloads and language models [1] - Extropic has raised $14.1 million in seed funding and has a team of around 20 engineers dedicated to research and development [2] Group 1: Company Background - Extropic was founded by Guillaume Weldon, who previously worked on quantum computing projects at Google [2] - The company has attracted talent from the quantum computing field, particularly those disillusioned by the limitations of quantum technology [2][3] - The CTO and co-founder Trevor McCaw also transitioned from Google's quantum computing project to join Extropic [3] Group 2: Technology and Development - The chip developed by Extropic features a unique array of square components designed to generate "programmable randomness," which can enhance simulations in various fields such as weather forecasting and financial modeling [1] - Current prototypes of thermodynamic hardware have been created in academic labs, including simple neural networks [1] - The anticipated delivery of Extropic's chips to early users later this year is expected to demonstrate value in high-tech trading and medical research through probabilistic simulation algorithms [3]