The economic account behind the AI boom: industry insiders break down the true cost of AI computing power

Golden Finance reports that on June 6th, according to CCTV Finance, with the application of artificial intelligence, the trends of raising "lobsters" and "horses" have sparked a nationwide craze, and the word Token is becoming more and more familiar to the public.
In March 2026, the National Data Administration designated the Chinese name for Token as "word element".
Also in March, China's daily call volume for word elements exceeded 140 trillion times, and each word element's generation relies on computing power, which in turn depends on electricity.
There is a saying that "60%–70% of AI computing power costs are electricity bills."
In response, Peng Li, co-founder, CTO, and chief hardware architect of Mu Xi Co., Ltd., stated in "Dialogue" that this refers to the proportion of operational and maintenance costs during operation, not the full lifecycle cost of the intelligent computing center.
A million-level intelligent computing center with an investment of about 1.5 billion yuan, in the entire investment cost, the proportion of electricity and maintenance costs does not exceed 20%.
However, during the operation phase, the main cost is indeed electricity bills, and at this point, energy efficiency becomes critical.
When discussing the requirements for choosing electricity partners, Peng Li said she mainly considers two aspects: one is cost; the other is "general, easy to use, stable, and reliable."
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned