By Katie Forster, with Luna Lin in Beijing
For weeks, the global tech industry has been waiting for a major AI launch from DeepSeek, seen as a benchmark for China’s progress in the fast-moving field.
It’s been more than a year since the startup put Chinese AI on the map in early 2025 with a low-cost chatbot that performed at a similar level to American rivals.
But despite reports and rumors of its imminent release, DeepSeek’s next-generation “V4” model is nowhere in sight.
Speculation is also rife over the geopolitical implications of which computer chips were chosen to train and power the new system: world-leading US designs or Chinese-made alternatives that the country is trying to develop.
“It’s important to know because at one level, it’s a signal of China’s AI self-sufficiency trajectory,” Wei Sun, chief AI analyst at Counterpoint Research, told AFP.
Tech news outlet The Information reported last week that the V4 could be used in the latest chips made by China’s Huawei.
Such a move would mark a milestone for China in its bid to defeat US restrictions on the export of high-end artificial intelligence 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 DeepSeek’s launch by tech giants including Alibaba, ByteDance and Tencent.
AFP contacted DeepSeek, Huawei, Alibaba, ByteDance and Tencent but none were available to comment.
‘Wake Up Call’
DeepSeek began life in 2023 as a side project of a hedge fund that had access to a cache of powerful Nvidia processors.
He grabbed attention in January 2025 with his deep-reasoning chatbot R1, which sent US tech stocks tumbling President Donald Trump called it a “wake-up call” for American firms.
The R1 was based on DeepSeek’s last major AI model, the V3, which was released in December 2024.
The company’s affordable and customizable 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 the V4 – said to be multimodal, meaning it can generate text, photos and video – could shake up US tech ratings again.
“I expect the upcoming release of DeepSeek V4 will not just be a software update; it will be a very capable, open-source model that handles massive context windows at a fraction of the cost,” he predicted.
But DeepSeek’s reputation as a company on the frontier of AI technology is also at stake.
Its models previously relied on Nvidia chips, so a move to partner with domestic chipmakers would require “significant re-engineering,” Wei said.
“This transition could slow down development cycles and introduce performance compromises, especially for the V4, a model expected to be more advanced.”
Training against inference
The US cites national security concerns as the reason for banning the export of Nvidia’s most powerful AI processors to China.
“The continued wait for DeepSeek V4 points to friction in scaling advanced models without unlimited access to high-end Nvidia hardware,” Wu said.
But some reports claim that DeepSeek overturned the ban to train V4 by using thousands of Nvidia’s high-end Blackwell chips, disassembled in third countries and smuggled into China.
Training AI models requires large amounts of computing power—much more than processing the AI’s generative questions, known as inference.
AFP has contacted DeepSeek for comment. Nvidia did not respond to a request for comment, but told The Information that it had not seen evidence of this and “such smuggling seems pointless.”
Another Chinese AI startup, Zhipu, in January unveiled an image generator it said was trained entirely on Huawei chips.
And Alibaba said this week it would open a new data center 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 change in the geopolitical technology landscape,” Wu said.
