U.S. President Joe Biden plans to sign an executive order this month restricting U.S. companies' investments in key areas of the Chinese economy, with the artificial intelligence and semiconductor sectors likely to face a total investment ban, Bloomberg reported. However, the Wall Street Journal reported on July 7 that these U.S. sanctions are prompting Chinese technology companies to accelerate the pace of research seeking to develop cutting-edge artificial intelligence without relying on the most advanced U.S. chips.

Internet expert Guo Tao told the Global Times on the 8th that the U.S. ban on high-end chips means the chip giant's high-end chips are losing China, the world's largest semiconductor market. In the long run, Chinese AI chip makers are expected to usher in a new opportunity to change lanes and overtake.

The Wall Street Journal says Chinese companies are working on technologies that might allow them to build cutting-edge AI performance with relatively few or less powerful semiconductors. Some Chinese companies are looking at ways to use a combination of different types of chips to avoid relying on any one type of hardware, while others are looking at combining three or four relatively low-performing chips to replace Nvidia's most advanced processors.

Chinese companies are also looking at using various software techniques to reduce the computational intensity of training large-scale AI models due to the lack of powerful chips to provide the computing power, and all are looking at ways to use the chips more effectively to keep down the surging cost of AI development.

Hong Yong, a think tank expert at China's 50 People's Forum on Digital and Real Integration, told the Global Times on August 8 that large models require a lot of computing resources to train and optimize, and that increased computing power can substantially improve the training speed and effectiveness of models. The U.S. restrictions on advanced chips will to some extent affect the research and development and production of Chinese companies. However, Chinese companies are also actively exploring alternative solutions to make up for the lack of high-end chips. Hong Yong, for example, said that Chinese companies can improve algorithm efficiency and accuracy by optimizing algorithms and reducing reliance on computing power; they can also improve algorithm efficiency and model performance by involving multiple parties in collaborative learning and sharing computing resources and model parameters, thus realizing distributed training of models.

According to Hong Yong, China already has many important achievements and excellent talents in the field of AI, which are important foundations for the development of China's AI industry. Although the U.S. restrictions on chips will bring certain challenges to China's AI industry, Chinese companies are capable of meeting these challenges through technological innovation and talent accumulation to maintain their competitiveness and momentum.