Huawei is accelerating the development of its next-generation artificial intelligence chip as it strengthens its position in the global AI semiconductor market and seeks to challenge Nvidia.
The Chinese technology giant has moved the planned launch of its Ascend 960DT AI chip forward to the first quarter of 2027, three quarters earlier than the previously announced third-quarter schedule.
The updated timeline was announced by Huawei Deputy Chairman and Rotating Chairman David Wang during the Huawei Connect 2026 conference in Shanghai.
A Huawei spokesperson also confirmed to TechCrunch that development of the Ascend 960 series is progressing ahead of schedule, with the company targeting roughly double the performance from one generation to the next.
The Ascend 960DT is part of Huawei’s broader effort to develop advanced AI computing hardware that can serve Chinese technology companies and AI developers.
Huawei is also preparing another chip in the same generation. The Ascend 960PR is currently scheduled for release in the third quarter of 2027, according to Reuters.
Beyond individual processors, Huawei is developing a broader computing architecture designed to connect large numbers of AI chips into unified systems.
The company has introduced its Peerium Computing Architecture and UnifiedBus technology to connect processors with memory, storage and networking components. Huawei says the approach is intended to support large AI systems used for both training and inference.
Huawei has said its Atlas 950 SuperPoD and SuperCluster systems are among the first products based on this architecture. The company says an Atlas 950 SuperCluster can connect up to 256,000 accelerator cards.
The accelerated Ascend 960DT schedule comes as Huawei continues expanding its AI hardware business despite restrictions affecting China’s access to advanced foreign semiconductor technology.
Reuters reported that demand for Huawei’s AI computing equipment in China currently exceeds the company’s production capacity, limiting its ability to expand internationally in the near term.
Huawei’s strategy also reflects the importance of building complete AI computing systems rather than relying solely on the performance of individual chips. By connecting large numbers of processors, the company aims to provide greater computing capacity for increasingly demanding AI models.
The development places Huawei and Nvidia at the center of an increasingly competitive AI semiconductor market. Nvidia continues to have a major software advantage through its CUDA ecosystem, while Huawei is expanding both its hardware and software ecosystem around Ascend processors.
The announcement also comes shortly before a scheduled September 24 meeting in Washington between US President Donald Trump and Chinese President Xi Jinping, adding to the broader technology and semiconductor context surrounding Huawei’s AI push.
With the Ascend 960DT now targeted for early 2027, Huawei is bringing its next major AI chip generation to market sooner than previously planned while continuing to develop large-scale computing systems capable of linking thousands of processors.
