Huawei Technologies, a leading Chinese tech company, introduced its latest AI computing device on Saturday during a three-day event in Shanghai. At the World Artificial Intelligence Conference (WAIC), the firm presented the CloudMatrix 384 system, which attracted a significant number of attendees gathered at its exhibit. The new system is aimed at competing with the top product of a major American chip manufacturer in the fast-growing field of artificial intelligence.
Since its announcement in April, the CloudMatrix 384 has captured the interest of experts around the world. Industry specialists now view it as a candidate to rival Nvidia’s advanced solution known as the GB200 NVL72. Dylan Patel, who heads the semiconductor research organization SemiAnalysis, commented in an April article that Huawei’s technology might outperform Nvidia’s offering, highlighting the potential impact on the national AI market.
The new system is engineered with 384 of Huawei’s latest 910C chips, demonstrating better performance on several metrics compared to Nvidia’s unit, which is built with 72 B200 chips, according to SemiAnalysis. This performance boost is attributed to the system’s design, which compensates for the lower output of individual chips by integrating a larger number within the overall architecture. The structure, described as a “supernode” configuration, allows the chips to connect with exceptionally high speeds.
Representatives present at the WAIC venue chose not to provide additional details regarding the CloudMatrix 384, and follow-up inquiries received no comment. In light of ongoing U.S. export restrictions, Huawei continues to be recognized as one of China’s foremost providers of chips critical to artificial intelligence research. Last month, Nvidia’s CEO, Jensen Huang, mentioned in remarks to Bloomberg that Huawei has been advancing rapidly, with the CloudMatrix serving as a clear example of its progress. In June, Huawei Cloud CEO Zhang Pingan confirmed that the system is now live on the company’s cloud network, marking an important step in broadening its AI capabilities.