DeepSeek V4 Model Set to Run on Huawei Chips, Report Reveals
DeepSeek is preparing to launch its V4 artificial intelligence model using chips developed by Huawei Technologies. According to recent reporting, major firms such as Alibaba, ByteDance, and Tencent have already placed large orders for these processors. As a result, demand signals strong confidence in the upcoming release.
Moreover, this move reflects months of planning behind the scenes. Instead of relying on long-dominant American hardware, DeepSeek has chosen to highlight locally produced technology. Consequently, the V4 launch is expected to showcase the growing capabilities of China’s semiconductor ecosystem.
Breaking From Industry Norms
Traditionally, AI developers provide early access to major chipmakers to optimize performance. However, DeepSeek took a different route by denying early access to Nvidia and AMD. Instead, it prioritized collaboration with Huawei Technologies and Cambricon Technologies.
As a result, Chinese chipmakers gained several weeks to fine-tune their processors. Furthermore, engineers worked closely to adapt parts of V4’s core code for better compatibility. This coordinated effort highlights a broader push toward technological self-reliance. At the same time, tightening U.S. export restrictions have accelerated this transition.
Rising Momentum and Global Implications
The strategy aligns with growing interest in Huawei’s latest AI processor, the 950PR. Recent testing has shown promising results, particularly for inference tasks. Therefore, companies are increasingly preparing to deploy the chip at scale.
In addition, production plans indicate significant growth ahead. Huawei aims to deliver hundreds of thousands of units this year, with full-scale shipments expected later. Consequently, the chip could become a key component in future AI deployments.
However, the decision also carries geopolitical implications. A senior U.S. official previously told Reuters that “DeepSeek’s model was trained using Nvidia’s advanced Blackwell chips, which are restricted for export to China”. Meanwhile, analysts noted, “DeepSeek will block Nvidia’s pre-model access and claim that V4 was trained based on Chinese AI semiconductors such as Huawei’s,” and added that the move “appears to be part of a broader strategy to push out Nvidia’s AI semiconductors from the Chinese market in the long term”.
Ultimately, this shift signals more than a technical adjustment. Instead, it underscores a strategic effort to reduce reliance on foreign technology while strengthening domestic innovation.
