By Eduardo Baptista and Ethan Wang
BEIJING, April 24 (Reuters) – DeepSeek, the Chinese artificial intelligence startup whose low-cost model stunned the world last year, on Friday launched a preview of its highly awaited new model adapted for Huawei chip technology, underlining China’s growing prowess in the sector.
The close collaboration with Huawei on the model, the V4, contrasts with DeepSeek’s past reliance on Nvidia’s chips, though the startup did not disclose which processors it used to train its latest model.
The pro version of the new model outperforms other open-source models in world-knowledge benchmarks, trailing only Google’s Gemini-Pro-3.1, which is a closed-source model, DeepSeek said.
The V4 also comes in a lower-cost flash version. Preview versions allow the company to incorporate real-world feedback and make changes ahead of a final product launch. DeepSeek did not provide a timeline for when the model is expected to be finalised.
DEEPSEEK AT CENTRE OF US-SINO AI TENSIONS
The launch of the preview comes one day after the White House accused China of stealing U.S. AI labs’ intellectual property on an industrial scale, threatening to strain relations ahead of a summit between U.S. and Chinese leaders next month.
DeepSeek has been at the centre of that controversy, accused by Washington of violating U.S. export controls by acquiring cutting-edge Nvidia chips to train its models. Anthropic and OpenAI have also said it improperly “distilled” their proprietary models.
Hangzhou-based DeepSeek has acknowledged the use of Nvidia chips but has not commented on whether those particular chips were subject to export bans. It has said that its V3 model used data naturally occurring and collected through web crawling and it had not intentionally used synthetic data generated by OpenAI.
The Chinese Embassy in Washington said it opposes “the baseless allegations,” adding that Beijing “attaches great importance to the protection of intellectual property rights.”
HUAWEI AND DEEPSEEK’S CLOSE COLLABORATION
Huawei, whose Ascend AI chip line is key to China’s efforts to reduce reliance on U.S. cutting-edge semiconductor technology, said on Friday it had worked closely with DeepSeek so the new V4 models could run across its full line of high-performance systems.
“The entire Ascend supernode product line now supports the DeepSeek V4 series models,” it said.
Washington began to restrict China’s access to advanced AI chips made by U.S. firms in 2022 and since then Beijing has accelerated its push to achieve technological self-sufficiency – a boon for domestic chipmakers like Huawei.
DeepSeek’s meteoric rise in early 2025 has also pushed low-cost, open-source models to the forefront of China’s AI ecosystem, spurring on a slew of domestic rivals.
Some of those rivals were on the back foot on Friday, however, as the release of the V4 sent their shares tumbling. Zhipu AI lost 9% and MiniMax slid 7%.
Owned by China’s High-Flyer Capital Management, DeepSeek is aiming to raise funds at a valuation exceeding $20 billion, according to a report by The Information this month, which also said that tech giants Alibaba and Tencent were in discussions to take stakes.
(Reporting by Ethan Wang and Eduardo Baptista; Editing by Edwina Gibbs)
Comments