BEIJING – Huawei Technologies is set to unveil new artificial intelligence technology this week in defiance of US export control curbs, advancing its ambition to replace Nvidia in China and compete on the global stage.
Huawei’s rotating chairman Wang Tao will present its upcoming Ascend 960 AI chips, to be commercially available in 2027, at the company’s annual summit in Shanghai on Sept 17.
Some critical details that analysts will watch out for include whether the new component will be powerful enough for training frontier AI models.
Huawei’s goal is to ensure its Ascend chips can operate every single AI model, in China and beyond, Chairman Guo Ping of the company’s supervisory board told new hires at an internal meeting earlier in September.
“We are closing the gap by innovating chip architectures,” Guo said. “Apple was not a leader in making processors at the beginning, but they delivered exceptional user experience through architecture redesigns and software optimisation.”
Already, Huawei has been making inroads at home and beyond. DeepSeek is planning to deploy at least 160,000 Huawei Ascend 950DT chips at a data centre it is building in Inner Mongolia to operate its models. Huawei is also exploring foreign markets including Malaysia and Egypt, Bloomberg News has reported.
The Chinese telecom-to-smartphone firm is spearheading Beijing’s drive for self-sufficiency in critical technologies in the face of yearslong US restrictions aimed at crippling China’s chipmaking aspirations.
It is also leading a Beijing-backed campaign to export the Chinese technology stack abroad at a time when the world’s two superpowers are vying for geopolitical dominance everywhere.
Ascend 960 will succeed Huawei’s current Ascend 950 family of AI accelerators, of which the 950DT favored by DeepSeek is equipped with faster, higher-capacity memory for demanding AI inference workloads. Another upgraded iteration, Ascend 970, is already planned for 2028, the Shenzhen-based company previously announced when launching a three-year campaign to overtake Nvidia.
Each new generation of chips will double the amount of computing power of its predecessor.
But Huawei’s AI chips are trailing Nvidia’s top-notch offerings. Frontier Chinese AI labs like DeepSeek are still largely relying on the US company’s chips to train their models, which they then operate using more domestic semiconductors, underscoring the challenge Beijing faces to wean itself completely off of western technology. At the same time, Nvidia’s most sophisticated silicon is still finding its way into China via smuggling, albeit in modest quantities.
A constraint in production capacity is also preventing Huawei from satisfying explosive demand for computing power in a market that Morgan Stanley estimates could reach 646 billion yuan (S$123 billion) by 2030. Nvidia’s most advanced Blackwell processors are not directly available in China. Huawei recently raised pricing for the Ascend 950DT chip by 60 per cent, citing tight component supply in communications with customers.
To overcome its disadvantages, Huawei has been trying to come up with innovative ways, including the recently announced LogicFolding technique, to improve its chip technology without the best gear.
Huawei has also introduced a so-called SuperPod structure that will soon allow as many as 100,000 Ascend chips to be linked in a cluster and enables faster data transmission via its UnifiedBus networking protocol. The company says its adaptations and networking can make up for shortcomings in single-chip performance relative to the likes of Nvidia.
“System-level competitiveness matters more than ever,” said Morgan Stanley analyst Charlie Chan in a recent research note. “The effective gap is narrowing through multi-die design, advanced packaging, rack-scale system architecture, optical networking, and software-hardware co-optimisation,” Chen wrote.
Huawei is also making some progress on other fronts. Ascend 950 chips are the first Huawei AI accelerators powered by memory chips it designs in-house, though the company has provided little production details.
The Shenzhen electronics giant and other Chinese AI chipmakers have been cut off from the latest AI memory products from South Korea’s SK Hynix and Samsung Electronics as well as US-based Micron Technology as part of US export controls. BLOOMBERG
