Huawei announced at its Huawei Connect conference that it has advanced the planned launch of its next-generation Ascend 960DT AI chip to the first quarter of 2027. The company had previously targeted the third quarter of 2027 for the chip's debut.
A Huawei spokesperson told TechCrunch that the Ascend 960DT is expected to be ready in Q1 2027, adding that Ascend 960 chips are launching ahead of schedule, with double the performance and annual improvements. David Wang, Huawei’s rotating and acting chairman, communicated the updated timeline at Huawei Connect on Thursday.
Peerium Computing Architecture and early systems
The announcement is part of Huawei’s effort to link hundreds of thousands, and eventually millions, of AI accelerators into very large computing systems. Huawei calls this approach the Peerium Computing Architecture, which relies on UnifiedBus — the company’s method for connecting processors, memory, storage and networking hardware.
According to Huawei's statement, Eric Xu, the company’s rotating chairman, said the Peerium architecture is being used to build larger AI computing systems intended for both training and inference. The Atlas 950 SuperPoD and Atlas 950 SuperCluster are the first systems built on the architecture; Huawei says an Atlas 950 SuperCluster can connect up to 256,000 accelerator cards.
Questions on scaling and geopolitical timing
Despite the accelerated chip timeline, analysts have raised questions about the scaling claims for Huawei’s broader systems. China tech analyst Rui Ma posted on X that Huawei had previously said the Atlas 960 SuperPoD would scale to 15,488 Ascend 960 chips, whereas this week’s announcement referenced a system with 4,096 chips. Ma noted that while the chip itself is arriving much earlier, the announced SuperPoD is considerably smaller than initially outlined.
The timing of the announcement also coincides with a tense geopolitical backdrop: it comes one week before U.S. President Donald Trump and Chinese President Xi Jinping are scheduled to meet on September 24 in Washington, DC.
Huawei has continued to push its chip ambitions despite U.S. restrictions on China’s access to advanced semiconductor technologies. Rui Ma said those restrictions are unlikely to stop China from pursuing its own semiconductor capabilities, arguing that the stakes for self-sufficiency are high.
On the broader debate about AI development, Donald Trump has resisted industry calls to slow progress on safety grounds, arguing the U.S. must preserve its lead over China. At the same time, Eric Xu told the Financial Times that China needs to accelerate AI development to catch up with the U.S., suggesting Chinese firms must advance further before they can fully assess and mitigate risks from more powerful AI systems.
What this implies
Moving the Ascend 960DT launch earlier signals Huawei’s intention to more aggressively contest major players in the AI accelerator market, notably Nvidia. The Peerium architecture and UnifiedBus aim to enable large, modular AI clusters; however, the practical achievable scale and operational deployment details remain under scrutiny. Analysts’ questions about SuperPoD sizing and the geopolitical context will shape how the market and policymakers respond to Huawei’s plans.
