Chongqing Targets Over 100,000 PFLOPS of Schedulable Intelligent Computing by 2028
The General Office of the Chongqing Municipal People's Government recently issued an action plan aiming for the city's schedulable intelligent computing scale to exceed 100,000 PFLOPS by 2028. The plan focuses on advancing token-driven data industries, building 20+ token innovation platforms, and fostering 40+ industry vertical models. This initiative significantly accelerates regional AI infrastructure and digital economy growth by aligning computing power with data spatial development. It establishes a localized framework for vertical model deployment across key sectors like smart manufacturing and urban governance. The action plan also outlines goals to build 100+ high-quality datasets and trusted data spaces, establish 10 characteristic data industrial parks, and nurture over 200 key data enterprises. It projects an average annual compound growth rate of over 10% for the city's data industry output value.
## BACKGROUND
Intelligent computing scale is measured in PFLOPS (petaflops) to quantify high-performance AI training and inference capabilities. Trusted data spaces and token-driven development refer to emerging concepts that treat structured data and tokens as core production factors in the agentic AI era.