The AI4SEI (full name: AI for Science, Engineering, and Industry) Industrial Innovation Consortium was recently officially inaugurated at China General Certification Center (CGC) in Beijing. Representatives from science‑technology, industry, investment, testing‑and‑certification and other sectors gathered at the founding conference. They conducted in‑depth discussions on topics including the AI4SEI innovation system, domestic chip validation testbed, construction and operation of ultra‑large‑scale computing‑power clusters, as well as technology‑oriented finance and industrial finance. A collaborative innovation organizational model integrating “government, industry, academia, research, capital and end‑users” was formally put into operation.

The consortium adopts a three‑tier structure consisting of a core layer, a close‑cooperation layer and an ecological layer. Centered on AI4SEI, it promotes deep integration of scientific‑technological innovation and industrial innovation. Taking project investment, construction and operation as the starting point and capital as the bond, it transforms sporadic exchange‑and‑cooperation activities into long‑term, stable, low‑cost and high‑efficiency partnerships. Through closed‑loop full‑chain operation for new technologies and emerging industries, the consortium will maximize internal innovation benefits and rapidly scale up external innovation spillovers.
Mr. Li Xinyu, Chairman of the Consortium and President of Beijing Academy of Science and Intelligence, delivered the keynote speech at the conference. Chairpersons of each special‑purpose committee presented their respective work plans, and on‑site members exchanged views on multiple topics. The meeting was presided over by Mr. Chen Xueyi, Secretary‑General of the Consortium, Chairman & President of Energy‑Computing Integration (Hami) Research Institute, and Industry Development Advisor of Beijing Academy of Science and Intelligence, who introduced the consortium’s organizational framework and operational mechanisms.
In his speech, Chairman Li Xinyu stated that AI has greatly unlocked productivity, yet new supporting infrastructure remains to be improved. Private enterprises will become the main innovation force in the AI4Science field. The state shall focus on key tasks that the market cannot accomplish and accelerate industrial development via innovative organizational mechanisms.
Qin Haiyan, Chairperson of the Power Group under the Consortium and Director of China General Certification Center, shared experience from wind‑power industry development. Drawing on years‑long practical experience of the wind‑power sector, CGC will undertake three core functions within the consortium: creating market space for emerging industries, building a solid quality‑verification foundation for high‑quality industrial development, and cooperating with local governments to drive project implementation.
Wu Jian, Chairperson of the Computing‑Power Group and Chairman of Tianjin Xinzhihuixiang Technology Co., Ltd., pointed out the existing imbalance between surging AI computing‑power demand and insufficient supply. His team will cooperate with the group led by Li Xinxi, Deputy General Manager of Beijing Xinli Semiconductor Co., Ltd., to carry out domestic chip validation work. Li Shuangxia, Chairperson of the Finance Group and Chairman of Beijing Mingde Investment Co., Ltd., introduced that the consortium will implement the integration of technology‑oriented finance and industrial finance to empower the real economy with capital. Li Chao, Chairperson of the Application Group and Vice‑President of Energy‑Computing Integration (Hami) Research Institute, predicted that the AI industry will shift from a triangular structure featuring large underlying infrastructure but small‑scale applications to an inverted‑triangle structure with explosive application growth. The conference also shared concepts and preliminary progress for the Science‑Education Computing‑Power Network and the Asia‑Europe Computing‑Power Network. Participants discussed Xinjiang’s strengths and roles in the national “East‑Data‑West‑Computing” initiative and the global expansion of Chinese AI, as well as approaches to boost cross‑regional collaboration between Beijing’s scientific research resources and Xinjiang’s industrial resources.
The consortium has defined three core missions:
Taking domestic chip validation as its business entry point, it focuses on domestic chip screening, hardware‑software adaptation, application adaptation, scenario‑oriented polishing and system optimization, to rapidly turn “usable” domestic chips into “high‑performance‑usable” products and achieve cost‑effectiveness and competitiveness through large‑scale deployment;
During the 15th Five‑Year Plan period, build 3‑5 clusters of 100 000‑card scale and a number of 1 000‑card and 10 000‑card clusters nationwide to form a hybrid physical‑virtual Science‑Education Computing‑Power Network. It will fully support the joint construction of a new‑generation super‑intelligent integrated 100 000‑card cluster by Beijing Academy of Science and Intelligence and Sugon;
Set up the AI4S Science‑Technology Innovation Fund and the AI4SEI Industrial Fund. It will coordinate three types of capital: long‑term science‑innovation capital (investing in early‑stage, small‑scale and hard‑technology projects), short‑term M&A capital (spanning from ICT to DAT), and large‑scale capital for computing‑power infrastructure. GW‑level computing‑power bases will be planned, with a target total investment of 300‑500 billion RMB for the 15th Five‑Year Plan period.
On July 10, 2026, the consortium held its preparatory meeting at the core node of the National Supercomputing Internet in Zhengzhou. Attendees witnessed the signing ceremony for the new‑generation 100 000‑card cluster jointly built by Beijing Academy of Science and Intelligence and Sugon, and discussed matters concerning the design, construction and operation of the 100 000‑card cluster. The consortium plans to convene its first closed‑door member meeting in Nanjing in October to discuss and arrange project construction work for 2027.