BEIJING, April 29 — Demand for Huawei’s Ascend 950 AI chips has surged following the release of DeepSeek’s V4 artificial intelligence model that runs on the Shenzhen-based tech firm’s chips, with major Chinese internet firms rushing to secure orders, three people familiar with the matter said.

China’s biggest internet firms including ByteDance, Tencent and Alibaba are reaching out to Huawei about new chip orders, said the sources, who are familiar with the procurement discussions.

Companies specialising in cloud computing and graphics processing unit (GPU) rental services are also scrambling to place orders, two of the sources added, without providing the names of the firms.

While the 950PR significantly outperforms Nvidia’s H20 chip — the most powerful chip Nvidia was permitted to sell in China until Beijing blocked its import last year — it still trails the American firm’s H200, a more advanced processor that has been caught up in regulatory limbo.

Despite US and Chinese approvals for exports, the H200 has yet to be shipped to China as Beijing and Washington remain at odds over the conditions governing its sale, providing an opportunity for Huawei to sell its semiconductors.

The 950PR represents a breakthrough for Huawei after years of struggling to win large orders from China’s tech sector. Customer testing of the chip went well earlier this year, with firms including ByteDance and Alibaba planning orders after samples were distributed in January, Reuters reported in March.

Huawei, ByteDance, Alibaba and Tencent did not respond to Reuters requests for comments.

DeepSeek frenzy

The scramble for Huawei’s chips underscores how DeepSeek’s V4 release last week has turbocharged demand for domestic Chinese AI hardware as US export controls continue to restrict access to Nvidia’s most advanced processors. It is also an endorsement of the performance of Huawei’s chips so far.

DeepSeek’s decision to optimize its V4 specifically for Huawei’s chips marks a strategic shift away from American semiconductor dependence and more towards China’s homegrown AI gear, which is a priority for Beijing as it seeks tech supremacy.

Last week, Huawei said its Ascend supernode infrastructure, built on the Ascend 950 series chips, would fully support DeepSeek V4 models and that the entire Ascend SuperNode product line had been adapted for V4 inference, referring to the process of using a trained AI model to answer queries and execute tasks.

Among Chinese chipmakers, Huawei’s Ascend 950 series — specifically the 950PR variant — is the only domestic chip to support a technique that processes AI calculations in a more compressed numerical format, allowing it to handle more computations per second at a lower cost.

Highlighting the rush for access, Alibaba Cloud’s Bailian platform made DeepSeek V4 available on the day that it was released, offering both the V4-Pro and V4-Flash variants with pricing matching DeepSeek’s official rates.

Tencent Cloud launched V4 preview services on its TokenHub platform on the same day, deploying the model on both domestic nodes and its Singapore international gateway to serve global users.

The rapid deployment by major cloud platforms means millions of users and developers can now access V4, sharply increasing the volume of AI queries that need to be processed — and with it, demand for the underlying chips.

Supply constraints persist

DeepSeek, which is offering developers a 75 per cent discount on its new model until May 5, said V4-Pro pricing could decline significantly in the second half of 2026 once Huawei’s Ascend 950 supernodes “ship at scale.”

However, the company acknowledged that constraints would persist until production ramps up, reflecting the tight supply of high-end homemade AI chips.

DeepSeek’s V4 includes two versions: V4-Pro with 1.6 trillion total parameters and V4-Flash with 284 billion parameters, both supporting a one-million-token context window. The models are available as open-source releases under the permissive MIT open-source licence, which allows companies to freely use, modify and commercialise the models.

However, output of the 950 is expected to fall short of demand due to US export restrictions on advanced chipmaking tools that prevent China from acquiring cutting-edge manufacturing equipment.

Huawei planned to ship around 750,000 units of the 950PR this year, with mass production beginning in April and full-scale shipments starting in the second half of 2026, according to people familiar with the plans. — Reuters

By Katie Forster, with Luna Lin in Beijing

For weeks now, the global tech industry has been waiting for a major artificial intelligence launch from DeepSeek, seen as a benchmark for China’s progress in the fast-moving field.

More than a year has passed since the startup put Chinese AI on the map in early 2025 with a low-cost chatbot that performed at a similar level to US rivals.

But despite reports and rumours about its imminent release, DeepSeek’s next-generation “V4” model is nowhere in sight.

Speculation is also swirling over the geopolitical implications of which computer chips were chosen to train and power the new system: world-leading US designs or made-in-China alternatives that the country is racing to develop.

“It’s important to know because at one level, it is a signal of China’s AI self-sufficiency trajectory,” Wei Sun, principal AI analyst at Counterpoint Research, told AFP.

Tech news outlet The Information reported last week that V4 can be run on the latest chips made by China’s Huawei.

Such a shift would mark a milestone for China in its bid to beat US restrictions on the export of top-of-the-range AI chips from Californian titan Nvidia to the country.

The report cited five people with direct knowledge of large orders for Huawei chips, made in preparation for the DeepSeek launch by tech giants including Alibaba, ByteDance and Tencent.

AFP contacted DeepSeek, Huawei, Alibaba, ByteDance and Tencent but none were able to comment.

‘Wake-up call’

DeepSeek started life in 2023 as a side project of a hedge fund that had access to a cache of powerful Nvidia processors.

It shot to attention in January 2025 with its R1 deep-reasoning chatbot, which sent US tech shares tumbling with President Donald Trump calling it a “wake-up call” for American firms.

R1 was based on DeepSeek’s last major AI model, V3, which was released in December 2024.

The company’s affordable, customisable AI tools have been widely adopted in China, and are also popular in emerging markets such as Southeast Asia and the Middle East.

Stephen Wu, founder of the Carthage Capital fund, told AFP that V4 — said to be multimodal, meaning it can generate text, pictures and video — could again shock US tech valuations.

“I expect the upcoming DeepSeek V4 release will not just be a software update; it will be a highly capable, open-source model that handles massive context windows at a fraction of the cost,” he predicted.

But DeepSeek’s reputation as a company at the frontier of AI technology is also at stake.

Its models previously relied on Nvidia chips, so a move to collaborate with domestic chipmakers would require “substantial re-engineering”, Wei said.

“That transition can slow development cycles and introduce performance trade-offs, especially for V4, a model expected to be state-of-the-art.”

Training vs inference

The US cites national security concerns as the reason for its export ban on Nvidia’s most powerful AI processors to China.

“The ongoing wait for DeepSeek V4 points to friction in scaling advanced models without unrestricted access to top-tier Nvidia hardware,” Wu said.

But some reports allege that DeepSeek skirted the ban to train V4 using thousands of Nvidia’s top-end Blackwell chips, dismantled in third countries and smuggled to China.

Training AI models requires huge amounts of computing power — much more than for processing generative AI queries, which is known as inference.

AFP has contacted DeepSeek for comment. Nvidia did not respond to a comment request but told The Information it had not seen evidence of this and “such smuggling seems farfetched”.

Another Chinese AI startup, Zhipu, in January unveiled an image generator that it said had been entirely trained on Huawei chips.

And Alibaba said this week it would open a new data centre for AI training and inference in southern China, powered by 10,000 of its own chips and operated by China Telecom.

As for DeepSeek, “if they have successfully trained V4 entirely on Huawei silicon, it signals a material shift in the geopolitical tech landscape”, Wu said.