Back
China's DeepSeek is developing its own AI chip, sources say

China's DeepSeek is developing its own AI chip, sources say

The reported move would expand DeepSeek beyond AI models into semiconductor design as it hires chip-design engineers and explores supply-chain partnerships for the project.

Fact Check
The Reuters exclusive directly supports every element of the claim: DeepSeek developing its own AI chip (inference-focused), expanding into semiconductor design, hiring chip-design engineers, and exploring supply-chain partnerships (chip-design, foundry, memory). Taipei Times independently corroborates the same details. The reporting is attributed to three sources familiar with the matter and is unconfirmed by DeepSeek itself, which is why probability is high but not certain. No conflicting evidence was found.
    Reference: 12
Summary

China's DeepSeek is developing its own AI chip and has recently stepped up hiring chip-design engineers for the effort, according to sources. The project began about a year ago and is focused on the AI inference stage, the part of AI deployment that handles running trained models in real-world use. Sources said DeepSeek is also discussing partnerships with chip design, foundry and memory companies, a sign the startup may be building out the supply chain needed to bring a custom processor to market. The reported push suggests DeepSeek is moving deeper into semiconductor design, a strategic area for AI developers because custom chips can help control computing costs, improve performance and reduce reliance on outside suppliers such as Nvidia and Huawei.

Terms & Concepts
  • AI chip: Semiconductor designed for artificial intelligence workloads
  • AI inference stage: Phase when a trained AI model is run to generate outputs in practical use
  • foundry: Company that manufactures chips designed by other firms