Huawei Accelerates Ascend 960DT AI Chip Launch to Q1 2027
The Chinese tech giant moves its next-generation AI processor timeline forward by six months as it builds systems to rival Nvidia's dominance.

Huawei pushes AI chip timeline forward
Huawei announced Thursday it will release its next-generation Ascend 960DT AI chip in the first quarter of 2027, moving the launch forward by six months from its original third-quarter target. The company revealed the accelerated timeline at its Huawei Connect conference, with acting chairman David Wang confirming the updated schedule.
The Ascend 960DT represents Huawei's latest effort to develop competitive AI processing hardware despite ongoing U.S. restrictions on China's access to advanced semiconductor technology. According to a company spokesperson who spoke with TechCrunch, the chip is expected to deliver double the performance of its predecessor while maintaining an annual advancement cadence.
New architecture for massive AI systems
Alongside the chip announcement, Huawei detailed its Peerium Computing Architecture, a framework designed to connect hundreds of thousands of AI processors into unified computing systems. The architecture relies on UnifiedBus, Huawei's proprietary technology for linking processors with memory, storage, and networking components.
Rotating chairman Eric Xu said the company is using this architecture to construct large-scale AI computing systems capable of handling both training and inference workloads. Huawei's Atlas 950 SuperPoD and SuperCluster represent the first implementations of this approach, with the company claiming an Atlas 950 SuperCluster can interconnect up to 256,000 accelerator cards.
Questions about system scale
The accelerated chip timeline has raised questions about the scope of Huawei's broader AI infrastructure plans. China technology analyst Rui Ma noted on X that previous Huawei statements indicated the Atlas 960 SuperPoD would scale to 15,488 Ascend 960 chips, while this week's announcement referenced a system with just 4,096 chips. The discrepancy suggests the chip itself is arriving significantly earlier than planned, but the corresponding system architecture may be launching at a smaller initial scale.
Why it matters
Huawei's accelerated chip development signals China's determination to build domestic AI infrastructure independent of Western suppliers, particularly Nvidia. The timing—just days before a scheduled meeting between U.S. President Trump and Chinese President Xi Jinping on September 24—underscores the geopolitical dimensions of AI chip competition. Ma argued that U.S. restrictions are unlikely to halt China's semiconductor ambitions, noting "the stakes for self-sufficiency are just too high at this point." The diverging approaches to AI development between the two nations—with Trump emphasizing maintaining U.S. leadership and Xu arguing China must accelerate to catch up—suggest the technology race will intensify regardless of diplomatic outcomes.
According to the Financial Times, Xu contended that Chinese companies need to advance their AI capabilities further before they can fully understand and address the risks associated with more powerful systems.
These details were first reported by TechCrunch.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call