Scholarlink

<p>The rapid growth of artificial intelligence is increasing electricity use in data centers while also enabling electricity savings through more efficient power system operation. Whether these savings can offset artificial intelligence's electricity consumption remains unclear. Here we develop a dynamic assessment method to measure net electricity saving, defined as the ratio of saved electricity to consumed electricity, for artificial intelligence in China’s power system. We examine nine scenarios that represent different levels of technology improvement and artificial intelligence adoption. The results show that the highest net electricity saving reaches 130.3% by 2040 under the best scenario. Net savings rise from 2040 to 2050, then decline by about 10% from 2050 to 2055 due to technology saturation. Electricity generation and electricity use dominate total savings, with strong regional variation, highlighting the importance of balanced deployment strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Net electricity savings from artificial intelligence depend on deployment efficiency in China’s power system

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月2日

查看全文
  • Kaile Zhou,
  • Ziwei Yang,
  • Rong Hu

摘要

The rapid growth of artificial intelligence is increasing electricity use in data centers while also enabling electricity savings through more efficient power system operation. Whether these savings can offset artificial intelligence's electricity consumption remains unclear. Here we develop a dynamic assessment method to measure net electricity saving, defined as the ratio of saved electricity to consumed electricity, for artificial intelligence in China’s power system. We examine nine scenarios that represent different levels of technology improvement and artificial intelligence adoption. The results show that the highest net electricity saving reaches 130.3% by 2040 under the best scenario. Net savings rise from 2040 to 2050, then decline by about 10% from 2050 to 2055 due to technology saturation. Electricity generation and electricity use dominate total savings, with strong regional variation, highlighting the importance of balanced deployment strategies.