Global Computing Power Price Index Launched at Financial Street Forum
CEIS launches global computing power price index with Jiuzhang Yunji as core co-builder.
According to Leiphone.com, the index is intended to support comparison and trading of computing power across specifications, platforms and regions, and to help build a new order in the global computing power market. China's computing power industry is expanding quickly as artificial intelligence demand rises, but the market still lacks a unified pricing benchmark and a widely recognized price measure, the report said.
Fang Lei, founder, chairman and CEO of Jiuzhang Yunji, said at the launch that the index gives AI-era productivity a "measurement tool." He said much computing power remains at the resource layer, leaving industry unable to calculate costs, finance unable to see output and local parks unable to carry out precise investment promotion. The index and platform, he said, turn heterogeneous scheduling, token output, utilization and delivery efficiency into observable, comparable and serviceable industry signals.
To make computing power understandable and priceable, the index uses Jiuzhang Yunji's "one unit of computing power" (DCU) as its basic measurement unit. The company compares it to a kilowatt-hour of electricity: different hardware models, cluster architectures and computing service specifications are converted into a standard unit. Computing power buyers need not understand hardware or cluster details, but can know how much model training or inference a DCU can support, according to Leiphone.com.
Jiuzhang Yunji's DCU-based practice in computing power measurement and pricing was selected for the Ministry of Industry and Information Technology's 2025 Typical Cases of Coordinated Development of New Information Infrastructure. As a co-builder of the index's measurement method, the company will take part in methodology design and data collection, and provide support for converting heterogeneous computing power and measuring effective computing power.
Cao Zhanzhong, executive dean of the Xinhua Index Research Institute, said at the event that the index will develop benchmark pricing rules for a "standardized computing power unit" and compile data from real global computing power markets to monitor price levels and capture price fluctuation signals. The platform will gather information on global computing power supply, demand, transactions and fulfillment, and provide scenarios for supply-demand matching, price publication and cross-border flow, according to Leiphone.com.