Right now, China's AI chip development is no match for the high-performance AI GPUs from NVIDIA and AMD. However, facing tightening export sanctions from the US government, Chinese authorities have rallied the collective efforts of the industry, government, and academia to find the opportunity to break through.
The "Chip Forum" of the 2023 World AI Conference (WAIC) was held on July 6 in Pudong, Shanghai. During the forum, the Chinese industry, government, and academia unveiled the Intelligent Leap Project (temporary translation). One of the focuses is to develop a general-purpose software/hardware interface to establish the standard for the AI industry.
Chinese chip companies such as Huawei's HiSilicon, Enflame, KunLunXin, FeiTeng, Hygon, MetaX, Smart Chip, iluvatar, ShenWei, Biren, Denglin, Fudan Microelectronics, Moore Threads, Riscure, Intellifusion, Corerain, Novauto, TheWake, Axera, and KiwiMoore have all endorsed this project.
Chinese media reports pointed out that several corporate representatives were present at the launch ceremony of the Intelligent Leap Project. These include EnFlame industrial standard manager Mei Jing-Qing, KunLunXin VP Sun Xiao-Si, iluvatar VP Ding Na, MetaX co-founder, CTO, and chief software architect Yang Jian, Denglin VP of government affairs Zhang Zhang, and Novauto business VP Tao Nan.
During the ceremony, Wu Jin-Cheng, director of the Shanghai Municipal Commission of Economy and Informatization, stated that providing market space and gathering strategic resources to promote the development of the AI chip industry is a key focus of the project.
The project will unite domestic and overseas industry, academia, and research institutions, leveraging advanced technology guidance to create a global smart chip ecosystem that is tailored for both domestic and international markets.
Additionally, the project will promote the creation of an evaluation system for AI chip performance testing, scenario testing, and overall assessment. It aims to establish AI industry standards, verify application scenarios, develop evaluation tools, recommend products, and publish results related to AI chips. Ultimately, the project will drive the R&D for a general-purpose software/hardware interface standard.
