Huawei Claims New AI Chip Beats Nvidia H20 by 2.8x

Huawei unveiled its new Atlas 350 AI accelerator card at the China Partner Conference, claiming the device delivers nearly 2.8 times better performance than Nvidia's H20 chip on certain inference workloads.

The Atlas 350 is powered by Huawei's Ascend 950PR processor and delivers 1.56 petaflops of computing power in FP4 precision format. The card carries up to 128 gigabytes of high-bandwidth memory and is specifically optimized for search recommendation, multimodal generation, and large language model inference workloads.

The H20 is Nvidia's chip engineered to comply with U.S. export restrictions while still serving Chinese customers. Huawei's claim of a roughly 2.8-times advantage positions the Atlas 350 directly against the most capable Nvidia hardware legally available in China, a market where AI infrastructure investment is surging.

Huawei also announced the FusionCube A1000 cabinet, designed to help small and medium-sized enterprises deploy AI quickly, extending the Atlas 350's reach beyond large cloud operators.

The launch fits the pattern Huawei has pursued since U.S. sanctions cut it off from advanced American chip technology, systematically building a self-sufficient AI hardware ecosystem for the Chinese market.

Performance claims have not yet been independently verified.