DeepSeek Launches V4 AI Model Optimized for Huawei Ascend Chips, Strengthening China’s Tech Self-Sufficiency

Date:

DeepSeek Launches V4 AI Model Optimized for Huawei Ascend Chips, Strengthening China’s Tech Self-Sufficiency

Beijing: DeepSeek, a Chinese startup known for its low-cost AI model that gained international attention last year, has unveiled a preview of its highly anticipated new model, V4, optimized for Huawei’s chip technology. This launch, which occurred on Friday, highlights China’s increasing self-reliance in the technology sector. According to DeepSeek, the Pro version of the V4 model surpasses other open-source models in global knowledge benchmarks, second only to Google’s proprietary Gemini-Pro-3.1.

The collaboration with Huawei for the V4 model marks a significant shift from DeepSeek’s previous dependence on Nvidia chips. He Hui, director of semiconductor research at consultancy Omdia, emphasized the importance of this development for China’s AI industry, noting that Huawei’s Ascend chips represent the country’s best domestically produced alternative to Nvidia. He stated that the support for DeepSeek’s V4 indicates that leading Chinese AI models can now operate on Chinese hardware.

Historically, many prominent AI models have relied on Nvidia’s chips for training and execution. Huawei confirmed that its chips played a role in the training process of the V4 model.

DeepSeek’s transition away from Nvidia underscores concerns raised by Nvidia CEO Jensen Huang and others regarding the potential loss of the developer ecosystem in China due to U.S. export controls and Beijing’s push for technological self-sufficiency. Huang remarked that the emergence of DeepSeek on Huawei’s platform would represent a troubling outcome for the United States.

Huawei and DeepSeek’s Close Collaboration

DeepSeek has faced allegations from Washington and U.S. competitors that its success is largely due to the improper use of American technology. This launch coincides with recent accusations from the White House, which alleged that China has been engaged in large-scale theft of intellectual property from U.S. AI laboratories.

It remains uncertain whether these issues will impact diplomatic relations, particularly with U.S. President Donald Trump scheduled to visit Beijing next month for discussions with Chinese leader Xi Jinping.

DeepSeek has acknowledged its use of Nvidia chips but has not clarified whether those specific chips were affected by export restrictions. The company has also stated that it has not intentionally utilized synthetic data generated by OpenAI.

Since 2022, the U.S. has imposed restrictions on China’s access to advanced AI chips produced by American companies, prompting Beijing to accelerate its efforts toward technological independence. Huawei has indicated that it collaborated closely with DeepSeek to ensure that the new V4 models are compatible with its entire range of high-performance Ascend systems.

Nvidia now faces increased challenges in regaining market share in China. Earlier this year, the Trump administration approved the sale of Nvidia’s powerful H200 chips to China, but sources have reported that shipments have been hindered by disagreements over sales terms in both the U.S. and China.

In pre-market trading, Nvidia’s shares remained stable, while Chinese chipmakers experienced gains, with Huahong Semiconductor and SMIC rising by 15% and 10%, respectively, amid expectations for broader adoption of domestic chips.

DeepSeek Now Faces Many Rivals

Numerous Western and some Asian governments have prohibited their institutions and officials from using DeepSeek, citing concerns over data privacy. Despite this, DeepSeek’s models have consistently ranked among the most utilized on international platforms that host open-source models.

In China, DeepSeek’s status as a national champion has diminished over the past year due to a surge of competitive offerings from domestic rivals. The company announced that the V4 model is particularly well-suited for AI agent tasks, which require more computational power than traditional chatbots.

The future success of the V4 remains uncertain. Analysts from Citi noted that the rapid evolution of AI models complicates the identification of a definitive market leader. They believe that the landscape is not a “winner-takes-all” scenario; rather, success will hinge on monetization, adoption rates, and sustainable revenue from services.

The release of the V4 has negatively impacted the stock prices of DeepSeek’s competitors, with Zhipu AI and MiniMax both experiencing declines of 9%.

In addition to the Pro version, the V4 will also be available in a lower-cost Flash version. These preview models allow DeepSeek to gather real-world feedback and make necessary adjustments before the final product launch. The company has not disclosed a timeline for the completion of the V4 model.

DeepSeek, backed by China’s High-Flyer Capital Management, is reportedly seeking to raise funds at a valuation exceeding $20 billion. Discussions are ongoing with technology giants Alibaba and Tencent regarding potential investments.

Source: www.emirates247.com

Read all the latest developments and breaking updates in the Latest News section.

Published on 2026-04-24 17:03:00 • By the Editorial Desk

Share post:

Subscribe

Popular

More like this
Related