Huawei Technologies Co. has unveiled a bold three-year plan to challenge Nvidia’s leadership in the artificial intelligence chip market, Bloomberg reported. While the company admitted its processors cannot yet match Nvidia’s raw power and speed, Huawei is betting on scale, connectivity, and government support to close the gap.
The announcement was made at Huawei’s annual Connect conference on Thursday by Rotating Chairman Eric Xu. The presentation, unusually public for the secretive Chinese tech giant, came just one day before U.S. President Donald Trump and Chinese leader Xi Jinping held their second phone call in four months.
As part of its vision, Huawei launched its next generation of Ascend AI chips alongside upgraded “SuperPod” systems. These advanced data center platforms integrate computing, storage, networking, software, and infrastructure management, similar to Nvidia’s approach.
Huawei’s strategy centers on linking up to 15,488 Ascend chips through its new UnifiedBus interconnect protocol. According to the company, this system enables data transfer between chips up to 62 times faster than Nvidia’s upcoming NVLink144 technology. For comparison, Nvidia’s current NVLink72 connects 72 Blackwell GPUs with 36 Grace CPUs.
The move marks a major shift in Huawei’s strategy. Since losing access to Taiwan Semiconductor Manufacturing Co. in 2020 due to U.S. sanctions, the company has operated with greater secrecy. This new roadmap signals a more open push to position Huawei as a true competitor in the global AI chip race.
