Guangdong Province Releases AI-related Policies

TapTechNews June 9th news, on June 6th, the General Office of the People's Government of Guangdong Province released the document 'Several Measures of Guangdong Province on the Enabling of AI in Thousands of Industries', mainly planning for the development of AI in Guangdong Province in 2025/2027.

Guangdong Province Releases AI-related Policies_0

TapTechNews learned from the file that Guangdong Province plans to have the provincial computing power reach 40 EFLOPS by 2025, and the scale of the AI core industry exceed 300 billion yuan; by 2027, the province's AI industry base will be further consolidated, and the computing power will exceed 60 EFLOPS, and the nation's leading algorithm system and computing power network system will basically be formed.

Specifically, the file mentioned that Guangdong Province will establish an AI chip ecosystem and build an intelligent perception industrial system. In 2027, an 'AI chip ecosystem' will be initially built, and the 'cale of the high-end intelligent sensor industry will be doubled'.

In terms of AI software, the file mentioned that the relevant departments will encourage enterprises to develop engine framework tools and develop intelligent operating systems, and at the same time launch self-controlled large-scale model products, strengthen the adaptability of AI software and hardware. By 2027, the scale of the intelligent software industry will reach 2.7 billion yuan, and the coverage rate of artificial intelligence self-controlled software will reach 50%.

The relevant file also proposed that 'high-quality Chinese data sets should be constructed'. The relevant departments will also jointly accelerate the supply of algorithm products and build large-scale model open source communities, and encourage enterprises to build high-quality Chinese language corpora databases for industries, and promote the collection, sharing and use of typical industry data. By 2027, more than 50 high-quality industry data sets will be constructed.

Likes