DeepSeek is reportedly making a bigger move toward Chinese-made AI chips as the company looks to reduce its dependence on advanced hardware from the United States.
The Chinese artificial intelligence company plans to increase its use of Huawei processors for training future AI models. Training is one of the most demanding parts of building an AI system because it requires large amounts of computing power and thousands of high-performance chips working together.
DeepSeek CEO Liang Wenfeng has reportedly told investors that using more domestic hardware for AI training is now an important priority for the company. Huawei could begin supplying DeepSeek with new chips suitable for AI training as early as the fourth quarter of 2026, with additional deliveries potentially continuing into early 2027.
The move is important because DeepSeek has historically relied heavily on Nvidia hardware for developing its most advanced AI models. Nvidia remains one of the world’s leading suppliers of processors used for artificial intelligence training.
However, U.S. restrictions have made it harder for Chinese technology companies to access some of Nvidia’s most advanced AI processors. That has encouraged companies in China to develop and adopt local alternatives.
Huawei has become one of the main companies trying to fill that gap. Its Ascend line of AI processors is being developed for large-scale AI workloads, including model training and inference.
DeepSeek has already been experimenting with Huawei hardware. The company has worked on adapting its models and software so they can operate efficiently on Chinese processors instead of depending entirely on Nvidia’s hardware and CUDA software ecosystem.
The latest development suggests DeepSeek wants to take that effort further by using Huawei chips during the actual training process, not only when running finished AI models.
Still, the transition is unlikely to happen overnight. DeepSeek continues to use Nvidia processors for some of its largest and most demanding models. Nvidia’s hardware and software ecosystem remains widely used because developers have spent years building AI systems around the company’s technology.
Huawei is also facing its own challenges. Demand for its AI processors in China is growing quickly, while production capacity remains limited. The company is working on new generations of Ascend processors and larger computing systems designed to connect huge numbers of chips together.
For DeepSeek, having another reliable source of AI computing power could become increasingly important as its models grow larger and require more resources to train.
The company has emerged as one of China’s best-known AI developers after gaining attention for producing powerful models while focusing heavily on efficiency and lower development costs.
Its growing relationship with Huawei also reflects a wider change in China’s technology industry. Chinese companies are investing more heavily in domestic processors, AI infrastructure and software as they try to build technology systems that are less dependent on U.S. suppliers.
Huawei chips may not completely replace Nvidia hardware at DeepSeek in the near term. Instead, the company appears to be building a more diversified computing strategy in which Chinese processors handle a growing share of its AI workloads.
If Huawei can continue improving the performance and supply of its AI chips, DeepSeek and other Chinese AI companies could gradually shift more model development toward domestic hardware.
The change could also increase competition in the global AI chip industry, where access to powerful processors has become one of the most important factors in developing advanced artificial intelligence.
