Arm公司的首席执行官雷内·哈斯近日发出警示，随着人工智能技术的不断发展，如果不提高能效，AI数据中心的电力消耗将成为一个日益严重的环境问题。据哈斯透露，目前美国AI数据中心的耗电量在全美电网中占比仅为4%，但这一数字在未来十年内可能将迎来巨幅增长。他预测，到2030年，AI数据中心的电力消耗占比可能会攀升至20%~25%。
哈斯特别提到了像ChatGPT这样的大型语言模型，指出此类应用在推动人工智能进步的同时，也加剧了能源消耗的问题。随着大模型的普及和应用范围的扩大，其所带来的能源挑战日益凸显。这一警告引发了业界对可持续发展和绿色计算的关注，如何在推进技术创新的同时，兼顾能源效率和环境保护，已成为亟待解决的关键议题。
哈斯的言论提醒人们，尽管人工智能带来了巨大的潜力和变革，但其背后的能源消耗问题不容忽视。未来，科技行业需要寻找更高效、更环保的解决方案，以应对AI数据中心不断增长的能源需求，同时确保全球能源结构的可持续性。
英语如下：
**News Title:** “Arm CEO warns: US AI data center power consumption could soar to 20%-25% of grid by 2030”
**Keywords:** Rising AI energy consumption, data center pressure, ChatGPT impact
**News Content:**
Arm CEO, René Haas, recently issued a warning that without increased energy efficiency, the power consumption of AI data centers could become a mounting environmental concern. According to Haas, these centers currently account for just 4% of the US power grid, but this figure is projected to see a significant surge over the next decade. He forecasts that by 2030, AI data centers could consume 20% to 25% of the country’s electricity.
Haas specifically mentioned large language models like ChatGPT, highlighting that while such applications drive AI advancements, they also contribute to escalating energy usage. As these models become more widespread and their applications broaden, the energy challenges they pose become more pronounced. This warning has sparked industry attention on sustainable development and green computing, with the need to balance technological innovation, energy efficiency, and environmental protection emerging as a critical issue.
Haas’s remarks underscore that while AI holds immense potential and transformative power, the energy consumption behind it cannot be overlooked. The tech industry now faces the task of finding more efficient and eco-friendly solutions to address the growing energy demands of AI data centers while ensuring the sustainability of the global energy infrastructure.
【来源】https://www.cls.cn/detail/1642509
Views: 1