HawkInsight

  • Contact Us
  • App
  • English

In defiance of the U.S. ban, China's purchase of banned Huida chips drives artificial intelligence progress!

Previously issued [Major incident] U.S. expands ban on Chinese chips, who can escape Huida, Intel and Supermicro?The report mentions that, following the initial chip ban in October 2022, the United States added a number of further restrictions in October 2023, including Huida's A100 and H100 chips, to limit China's technology and trade development.。However, according to the public tender file, there are loopholes in the underground market, making the ban useless, multiple institutions can still obtain chips, for which the United States is looking for solutions that can effectively improve the loopholes.。

无视美国禁令,中国购买被禁辉达晶片推动人工智慧进展!

According to a tender file reviewed by Reuters, the Chinese military, state-owned artificial intelligence research institutes and universities have purchased small batches of Nvidia (NVDA) semiconductors that have been banned from export by the United States in the past year.。Despite the U.S. ban, public filings show that dozens of Chinese entities have purchased Pfizer chips since the restrictions were imposed, including its A100 and H100 chips.。

These chips are considered more effective for artificial intelligence work, but due to the lack of good alternatives, Chinese companies still buy these banned chips。Purchasers include some US export-restricted entities such as Harbin Institute of Technology and University of Electronic Science and Technology of China。In addition, Reuters has not been able to determine how these vendors obtained these wafers.。

Although the United States is working to close export restriction loopholes, an underground market for such chips has formed in China。Pfizer says it complies with all applicable export control laws。Experts believe that U.S. export restrictions are difficult to completely stop, but the goal is to hinder China's construction of large advanced chip clusters, making it difficult to train artificial intelligence systems.。

The review also revealed that more than 100 tender files, including state-owned entities, procured A100 chips, while dozens of files showed A800 purchases.。Buyers include Tsinghua University, military entities, etc., most uses are artificial intelligence。Although the number of purchases is relatively small, it is enough to run a machine learning task。

Disclaimer: The views in this article are from the original author and do not represent the views or position of Hawk Insight. The content of the article is for reference, communication and learning only, and does not constitute investment advice. If it involves copyright issues, please contact us for deletion.