As the U.S. contemplates stricter trade measures to prevent advanced chip technology from being supplied to China, Nvidia, a prominent chip manufacturer based in the U.S., is reportedly developing a modified version of its latest artificial intelligence chips to adhere to these regulations.
According to sources cited by Reuters, Nvidia is creating a variant of its new Blackwell AI chips specifically for the Chinese market. The company plans to collaborate with a local distribution partner, Inspur, to introduce and sell this chip, provisionally named the “B20,” in China.
Reports indicate that the B20 is anticipated to begin shipping in the second quarter of 2025, although Nvidia has opted not to comment on the situation.
Nvidia currently has three chips tailored to meet U.S. export regulations, including the H20. To boost competitiveness against local rival Huawei, Nvidia has lowered prices for the H20 amidst lackluster sales. However, recent reports suggest that sales of the H20 are on the rise, with projections estimating that Nvidia will sell over one million of these chips in China this year, generating approximately $12 billion in revenue, despite existing U.S. trade bans. This expected volume is nearly double what Huawei anticipates for its Ascend 910B chip, according to data from SemiAnalysis.
On the other hand, analysts from Jeffries have cautioned that Nvidia’s H20 chips may face further jeopardy under additional U.S. trade regulations. They noted that during the upcoming annual review of semiconductor export controls in October, it is likely that the H20 will be prohibited from being sold to China. Potential methods for enforcement of such a ban could include a product-specific ban, adjustments to the computing power limit, and/or restrictions on memory capacity.
Moreover, the U.S. could expand its export restrictions to include chips sold to other nations in the region, such as Malaysia, Indonesia, and Thailand, or extend these controls to Chinese companies operating abroad, although the latter would present more implementation challenges, analysts reported.