Eager to launch the world’s most powerful computing clusters, Huawei said it wants to be a critical part of China’s mission to rely less on international semiconductor suppliers like Nvidia.
Speaking at Huawei Connect 2025 in Shanghai, Huawei’s deputy chairman of the board and rotating chairman Eric Xu said “computing power is – and will continue to be – key to AI. This is especially true in China.”
He explained that Huawei’s goal is to sustainably meet long-term computing demand by building SuperPoDs and SuperClusters with the semiconductor manufacturing process nodes that are practically available to China.
During the keynote, Xu revealed Huawei’s new SuperPoDs and SuperClusters. The news comes as the company broke its previous secrecy about its chips business, with detailed timelines for its Ascend AI chips and Kunpeng server chips.
Huawei describes a SuperPoD is a single logical machine made up of multiple physical machines that can learn, think and reason as one. During the event, Xu unveiled the company’s newest SuperPoD products: the Atlas 950 SuperPoD (with 8,192 Ascend NPUs) and the Atlas 960 SuperPoD (with 15,488 Ascend NPUs).
These two SuperPoDs are designed to deliver industry-leading performance across multiple key metrics, including the number of NPUs, total computing power, memory capacity and interconnect bandwidth.
Based on publicly announced product roadmaps from across the industry, Huawei’s SuperPoDs are reported to be the most powerful in the world.
During his keynote speech, Xu also announced the Atlas 950 SuperCluster (with over 500,000 Ascend NPUs) and Atlas 960 SuperCluster (with over one million Ascend NPUs), which are large-scale computing clusters comprised of multiple Huawei SuperPoDs. According to the company, these are also poised to outperform all other computing clusters on the market.
With this new technology, Xu said Huawei now has what it takes to provide abundant computing power for ongoing, rapid advancements in AI now and in the future.
Huawei has also introduced the TaiShan 950 SuperPoD, the world’s first general-purpose SuperPoD. Combined with Huawei’s distributed GaussDB, the company said the technology could serve as a viable alternative to mainframes and mid-range computers and Exadata database servers.
Honing its expertise for decades and combining a range of systems innovations, Huawei said it has overcome bottleneck challenges for large-scale AI computing infrastructure – including the physical limitations of existing cable technology to link up chips and SuperPoDs over long distances, whilst also being able to maintain low-latency connection.
The result is UnifiedBus, a groundbreaking interconnect protocol for SuperPoDs. Xu also released the technical specifications for UnifiedBus 2.0 at the event, in the hopes that industry partners will adopt this protocol to develop more UnifiedBus-based products and components and jointly create an open UnifiedBus ecosystem.
“SuperPoDs and SuperClusters powered by UnifiedBus are our answer to surging demand for computing, both today and tomorrow,” Xu explained. “Our goal is to keep pushing advancements in AI to create greater value.”
Huawei first announced plans to enter the chip market in 2018, but then went dark after the US sanctioned the company in 2019, citing issues of national security – which the company denied.
Since then, it is clear that Huawei has been pushing to become a Chinese leader in the semiconductor manufacturing industry.
In recent weeks, China has ramped up its efforts to target Nvidia, the world’s dominant AI chipmaker, in a push to prioritise domestic chip manufacturing. As reported by Reuters, the country has said that Nvidia has a monopoly on the market and the internet regulator has ordered top technology firms to stop purchasing Nvidia AI chips.
RELATED STORIES
