State-owned Assets Supervision and Administration Commission of the State Council: State-owned central enterprises (SOEs) must continue to tackle original technologies such as foundational algorithms, physical AI, intelligent computing chips, and quantum communications.

PANews July 18 report: According to Shanghai Securities News, Pan Xiaogang, deputy director of the State-owned Assets Supervision and Administration Commission (SASAC), said at the China Telecom artificial intelligence (AI) ecology forum held during WAIC2026 that: “State-owned central enterprises must stay abreast of the frontiers of the industry, continuously iterate on foundational large models, tackle foundational algorithms, physical AI, intelligent computing chips, quantum communication, and other original technologies; adhere to the principle of ‘state-owned models, state-owned chips, and state-owned applications’; and improve a fully independent end-to-end ecosystem spanning both software and hardware.” It was reported that SASAC has deepened its “AI+” special initiative for central enterprises, introduced targeted support policies, and for 3 consecutive years has carried out tiered and categorized talent intensive training. It is driving central enterprises to cultivate open, priority focus in more than 1,200 scenarios across key industries, actively adapting to domestically produced chips across all categories, has put into production 5 ten-thousand-card computing clusters, and is accelerating empowerment across all sectors and industries. Pan Xiaogang said that state-owned central enterprises should increase investment efforts, expand the scale of self-built and centrally administered intelligent computing, enhance cross-domain heterogeneous computing scheduling capabilities, accelerate the coordinated rollout and promotion of computing power, establish direct connections with green electricity, and provide compute power services to the public by offering equitable support for applications.