DeepSeek develops own AI chips to increase independence
DeepSeek, a prominent player in AI development, has commenced the design of its own AI chips to reduce reliance on external suppliers such as Nvidia and Huawei.

Vad har hänt
DeepSeek, known for its models DeepSeek-V2 and DeepSeek Coder, has announced that it is developing its own AI chips. This strategic initiative aims to eliminate the company's reliance on existing AI hardware suppliers, such as Nvidia and Huawei. By designing its own silicon hardware, DeepSeek intends to achieve greater control over its technological infrastructure and supply chain.
Key facts
| Företag | DeepSeek |
|---|---|
| Berörda leverantörer | Nvidia, Huawei |
| AI-modeller | DeepSeek-V2, DeepSeek Coder |
Varför det spelar roll
The decision to develop proprietary AI hardware is significant in a market dominated by a few chip manufacturers. By integrating hardware and software development, DeepSeek can potentially optimise its AI models more effectively. This could lead to improved performance, lower costs, and increased pace of innovation, strengthening their long-term competitiveness.
Vem påverkas
This primarily affects DeepSeek and its partners, who are expected to benefit from increased efficiency and reduced dependency. Additionally, Nvidia and Huawei are affected, as they may face reduced demand from DeepSeek. More broadly, the AI industry is impacted as more players may be inspired to consider proprietary hardware solutions, potentially leading to increased competition and innovation in chip manufacturing.
EU-status
Not applicable to EU status.
Mer att veta
The initiative to develop proprietary chips is symptomatic of a growing trend among major AI companies to exert greater control over their entire technology stack, from base hardware to advanced software models.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Länken öppnar i nytt fönster och leder till utgivarens egen sida.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.