The US ban on Nvidia and AMD selling some advanced chips to China is set to crimp the ability of Chinese tech companies in areas such as facial recognition and data centres, according to Chinese analysts and industry insiders. The ban on selected chips, which China is still unable to make itself, marked an escalation from current trade sanctions targeted at specific companies and sales restrictions on advanced chipmaking equipment. The impact of the new ban could be widespread for downstream users of chips in applications such as the metaverse, artificial intelligence, cloud computing and even facial recognition, industry experts warned. Nvidia said that US officials instructed it to stop selling its powerful A100 and H100 graphics processing units (GPU) to customers in China, while AMD said it has received new license requirements that will prevent its MI250 artificial intelligence chips from being exported to China, Reuters reported. The US Department of Commerce has not officially announced the new restrictions. China’s semiconductor industry now faces weak demand for low-end chips Chinese foreign ministry spokesperson Wang Wenbin said on Thursday that the US ban was “typical technology hegemonism” and an abuse of the national security concept. The US “is attempting to take advantage its own technology edge to constrain and depress development of emerging markets and developing countries … China firmly opposes it,” Wang said at a regular press briefing. China’s Ministry of Commerce spokesperson Shu Yuting said it noted the report, adding that China firmly opposed such restrictions as they would “hurt not only legitimate rights of Chinese companies but also interests of American businesses”. News of the latest US chip restrictions also stirred debate in China’s chip industry. Wang Xiaolong, research head at Shanghai-based semiconductor consultancy ICWise, said the new US export controls were a worrying sign for China as it still must rely on US firms for many necessary products and components. “In the long-term, China’s development in technology will be affected if export controls further escalate to more semiconductors such as CPU and FPGA chips,” Wang said, referring to two other advanced types of chips made by US firms. Nvidia said the new ban could mean the loss of up to US$400 million in sales bookings of the affected chips to China this quarter. It is not clear how many Chinese end users of the Nvidia and AMD GPU chips would be affected, but Wang pointed out that China’s large data centres require powerful GPUs as they have an edge in handling large troves of data in real time. An industry professional, who declined to be named as he was not authorised to speak to media, said Chinese companies without sufficient inventories of the targeted chips would struggle to find alternatives, and would have to resort to using less efficient versions. Zhu Jing, a deputy secretary general of the Beijing Semiconductor Industry Association, was quoted by local paper the Economic Observer as saying on Thursday that the latest US move could hinder China’s ability in advanced computing and affect its big data centre projects, including the eight national computing hubs that are part of a state plan to boost computing resources in Western China. The ban is part of Washington’s efforts to deny China access to advanced US chip technologies over fears they could be used by the Chinese military. On top of targeted sanctions against a growing list of Chinese entities, including national telecoms champion Huawei Technologies Co, US lawmakers passed the CHIPS and Science Act, which restricts American chip manufacturers from investing in advanced manufacturing facilities in China for a decade if they accept federal subsidies. New US tech export controls block China from achieving semiconductor ambitions The US is also pressing Dutch firm ASML to restrict Chinese sales of its high-end EUV lithography systems, which are used for the most advanced types of chips. Separately, Washington is backing a Chip4 Alliance with Japan, South Korea and Taiwan, a move that Beijing sees as an attempt to marginalise China’s role in global supply chains. China has been making significant progress in GPU research and development, but the country is still years behind AMD and Nvidia in advanced versions. Jingjia Micro, one of China’s best hopes for domestic GPUs, is still in its infancy, with total revenues of just 544 million yuan (US$78.8 million) in the first half of 2022, compared to Nvidia’s US$15 billion in the six-month period ending July. An employee at a large semiconductor design company, who requested anonymity due to the sensitive nature of the matter, said many Chinese GPU products coming out of labs have not been tested in the commercial market.