On December 6th, NVIDIA CEO Huang Renxun stated during his visit to Singapore that he will provide the Chinese market with a new set of products that comply with the latest regulations of the US government. Huang Renxun stated that he is working closely with the US government to ensure that new chips targeting the Chinese market comply with export restrictions. He added that Huawei is one of the "very strong" competitors in the competition to produce the "best" artificial intelligence chips.
Just a few days ago, on December 2nd, US Secretary of Commerce Raymond stated at the annual Reagan National Defense Forum in Simi Valley, California that China cannot be allowed to acquire cutting-edge chips.
Raymond pointed out Nvidia and warned that if the company redesigns chips around specific cutting lines to enable China to implement artificial intelligence projects, "I will take control measures the next day." Raymond said that in addition to the current $200 million budget, her department needs more resources to control the export of American chips.
On December 4th, Chinese Foreign Ministry spokesperson Wang Wenbin responded to Raymond's remarks by saying: President Biden has stated that he has no intention of obstructing China's economic development or hindering China's technological progress. However, the statements made by US officials are contradictory, making it difficult to trust China and the international community, and also exposing the deep-rooted Cold War mentality and hegemonic mentality of some US officials. China never gambles on the United States losing, nor does it intend to challenge or replace the United States. The US should establish a correct understanding of China and work with China to effectively implement the San Francisco summit between China and the US dollar The important consensus reached during the meeting is to stop viewing China as a 'hypothetical enemy'
Previously, the US Department of Commerce's Bureau of Industry and Security (BIS) 1017 new regulations came into effect on November 16th. Nvidia's "special edition" A800 and H800 chips launched for the Chinese market were subject to export compliance restrictions and could not be sold to Chinese customers. However, Nvidia continues to follow US policy adjustments to sell castrated chips to China. Shortly before the new regulations came into effect, it was revealed that Nvidia would launch three AI chips targeting the Chinese market, including H20, L20, and L2. But later Nvidia stated that the release of H20 will be postponed to the first quarter of next year.
On November 22, Nvidia released its financial report for the third quarter of the 2024 fiscal year (as of October 2023). According to the financial report, Nvidia achieved a revenue of 18.12 billion US dollars during the period, an increase of 206% year-on-year; The net profit was 9.243 billion US dollars, a year-on-year increase of 1259%.
Nvidia warned in its financial report that it expects a significant decline in sales in China in the fourth quarter due to the implementation of new regulations in the United States. Huang Renxun pointed out that China contributes about 20% of Nvidia's revenue, but it is difficult to predict how much the situation will change in the future with the new export restrictions in the United States.
According to the IT Times, since the US imposed an AI chip export ban on October 24th, the prices of domestic computing servers have skyrocketed. A dealer said, "The prices of H800 and A800 servers (not equivalent to the rental price of computing power) are very high now, with one unit soaring to around 3.5 million and 2 million, which is twice as high as before."
According to a report by market research firm IDC, with government support, local cloud AI acceleration chip manufacturers in China are also growing rapidly. In the first half of 2023, China's AI servers used 500000 locally purchased/developed AI accelerator chips. This quantity accounts for 10% of the entire server market in China. This proportion is clearly expected to soar in 2024.
At present, domestic AI chip manufacturers mainly include Cambrian, Jingjiawei, Yuntian Lifei, Hengshuo Shares, Haiguang Information, Fudan Microelectronics, Anlu Technology, Lanqi Technology, Hangyu Micro, Guoxin Technology, Ziguang Guowei, Guoke Micro, etc. Among them, Cambrian's ASIC chips are the leading domestic AI chip manufacturers, while Jingjiawei's GPU chips are the leading domestic GPU manufacturers.