Alibaba and Baidu Shift to In-House AI Chips

Sep 12, 2025
Alibaba and Baidu have started using their own chips for AI model training, partly replacing Nvidia's, as reported by The Information.

Alibaba Group and Baidu, Inc. have begun using their own internally designed chips to train AI models, partly replacing those made by NVIDIA, according to The Information. This move marks a significant shift in the Chinese tech landscape, where companies have traditionally relied on NVIDIA's powerful processors for AI development.

Alibaba has been utilizing its Zhenwu processing unit for smaller AI models since early this year, while Baidu is experimenting with its Kunlun P800 chip for training new versions of its Ernie AI model. Despite this shift, both companies continue to use NVIDIA chips for their most advanced models due to performance advantages.

The transition to in-house chips comes amid increasing U.S. export restrictions on advanced AI chips to China and pressure from the Chinese government to adopt homegrown technology. This strategic move by Alibaba and Baidu reflects a broader trend of Chinese companies seeking to reduce reliance on Western semiconductor technology.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.