China’s DeepSeek Accelerates AI Chip Development to Reduce Reliance on Nvidia and Huawei

Chinese startup DeepSeek is reportedly advancing its efforts to develop an in-house AI chip, a move that could significantly lessen its dependency on Nvidia and Huawei technologies. This initiative is particularly noteworthy as DeepSeek has relied on these companies’ chips to train and operate its widely acclaimed AI models.

The new chip is intended for inference, which is the phase of AI computing where trained models generate responses for users, rather than for the initial training of those models. If successful, this venture into semiconductor development would represent a strategic pivot for DeepSeek, a company recognized in China as a leader in AI innovation. This shift could also pose additional challenges for Huawei, a major player in the tech industry.

DeepSeek gained international attention over a year ago with the release of two highly efficient AI models that achieved viral status, surprising many observers in Silicon Valley and Washington. Historically, the company has focused on breakthroughs in AI models rather than the commercialization of its technology.

Despite Huawei’s offerings still falling short of Nvidia’s most advanced chips, U.S. export restrictions have allowed Huawei to capture approximately half of the $50 billion domestic AI chip market, supplying DeepSeek and other leading firms. However, Huawei’s market dominance is beginning to wane as competitors like Alibaba and Baidu develop their own AI chips and expand their market share.

Early Stages of Chip Development

DeepSeek’s foray into chip development is still in its infancy. The company has begun reaching out to external partners and engaging in discussions with chip-design, foundry, and memory companies. This initiative reportedly commenced about a year ago. Additionally, the Hangzhou-based firm has ramped up its hiring of chip-design engineers in recent months, although these recruitment efforts have been conducted privately, without public job postings.

Sources familiar with the matter have chosen to remain anonymous due to the sensitive nature of the information. Despite its status as a flagship of China’s AI ambitions, DeepSeek has maintained a low profile and has not responded to requests for comment.

Aligning with Global Trends

By developing its own chip, DeepSeek would align itself with other global AI developers who are seeking greater control over the hardware that powers their models, thereby reducing reliance on Nvidia. Recently, OpenAI introduced Jalapeno, its first custom inference chip developed in collaboration with Broadcom. Meanwhile, Anthropic is also considering the development of its own AI chips.

For DeepSeek, this initiative carries additional strategic importance. U.S. export controls prevent Chinese companies from acquiring Nvidia’s most advanced chips, prompting Beijing to encourage its technology leaders to create domestic alternatives. DeepSeek’s founder, Liang Wenfeng, acknowledged in a rare 2024 interview that these export controls pose challenges for the company.

Historically, DeepSeek has utilized both Nvidia and Huawei chips. The foundation model for its R1 reasoning model, which notably impacted U.S. tech stocks in January 2025, was trained on Nvidia’s H800 chip, which was banned by Washington in late 2023. Since then, DeepSeek has increasingly relied on Huawei’s technology. In April, the company launched its V4 model, which was adapted for Huawei’s Ascend chips. Huawei confirmed that its processors were partially used in training V4-Flash, a lighter version of the model. Following the launch, demand for Huawei’s Ascend 950 chips surged among Chinese tech conglomerates.

Addressing Inference Demand

DeepSeek’s planned inference chip aims to target the rapidly growing segment of AI computing demand. As AI applications proliferate, the industry’s focus is shifting from training models to running them. This shift necessitates specialized chips that are often more cost-effective and energy-efficient than traditional general-purpose GPUs.

However, the path to success is fraught with challenges. Designing a competitive AI chip typically requires years of development and substantial financial investment. Manufacturing presents another obstacle, as U.S. restrictions limit Chinese designers’ access to the most advanced overseas foundries. Additionally, separate U.S. regulations have restricted China’s access to high-bandwidth memory, a crucial component for AI inference chips.

DeepSeek’s push into chip development coincides with its first acceptance of external investment. The company is reportedly set to raise $7 billion in an initial funding round, valuing it between $52 billion and $59 billion, marking a significant departure from its previous strategy of eschewing outside capital.

Source: www.emirates247.com

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

Published on 2026-07-07 14:44:00 • By the Editorial Desk