Nvidia has informed its customers in China about a postponement in the launch of a new artificial intelligence chip it designed to comply with U.S. export rules. The chip in question, the H20, is the most potent among three chips created by Nvidia for the Chinese market to comply with the latest U.S. export restrictions. This delay might pose challenges for Nvidia in maintaining its market share in China, particularly against local competitors like Huawei. Originally expected to be launched around November 16, the release of the H20 has now been rescheduled for the first quarter of the upcoming year, potentially in February or March. This report was originally covered by Reuters.
It’s indicated that server manufacturers were facing challenges in integrating the H20 chip, leading to its delay. Alongside the H20, Nvidia had been developing two other chips, the L20 and L2, to comply with the updated U.S. export regulations. The sources clarified that the L20 is on track and will be launched as scheduled, but information on the status of the L2 remains unavailable. Nvidia sees these chips as crucial for maintaining its presence in the Chinese market after facing restrictions on shipping products following the tightened U.S. export rules. The A800 and H800, introduced as substitutes for Chinese customers in November 2022, were among the products affected by the U.S. export restrictions.
The H20, L20, and L2 chips incorporate many of Nvidia's latest AI features, although some computing power measures have been scaled back to comply with the new U.S. regulations, as reported by chip industry newsletter SemiAnalysis. The export constraints have created an opportunity for competitors like Huawei to secure orders that might have otherwise gone to Nvidia, whose GPUs dominate the AI market. Notably, Chinese tech giant Baidu reportedly placed a substantial order for Huawei AI chips, anticipating future challenges in purchasing from Nvidia due to the U.S. export restrictions.
