Alibaba, Baidu start using own chips to train AI models: sources

Alibaba and Baidu have begun using their own chips to train AI models, partly reducing reliance on Nvidia, according to a report by The Information.

Alibaba has deployed its in-house chips for smaller AI models since early 2025, while Baidu is testing its Kunlun P800 chip for training new versions of its Ernie AI model.

Both companies are major technology firms in China with significant AI operations.

The shift comes as US export controls limit Chinese access to advanced AI chips, prompting local firms to increase development of domestic alternatives.

The report said Alibaba and Baidu still use Nvidia chips for their most advanced AI systems, and have not fully replaced foreign technology.

🔗 Source: Reuters

🧠 Food for thought

Implications, context, and why it matters.

Chinese tech giants’ chip transition reflects years of strategic preparation

Alibaba and Baidu’s move to use internal chips represents the practical implementation of China’s long-term AI self-sufficiency strategy that began over a decade ago.
Chinese cities are now targeting 70% AI chip self-sufficiency by 2027, with local governments actively promoting domestic chip industries to reduce dependency on foreign technology1.
This transition builds on years of substantial investment, including early efforts like Cambricon Technologies, which raised $100 million in funding and aimed to capture a significant portion of the Chinese AI chip market2.
Xi Jinping’s call for a nationwide effort to build an independent AI ecosystem has mobilized both state resources and private sector innovation, with companies emerging as competitive alternatives to Western AI models3.
The Chinese government has launched significant funding initiatives, including a national AI investment fund, demonstrating coordinated support across multiple levels of government3.

Export restrictions accelerate rather than prevent Chinese technological independence

U.S. export controls have inadvertently strengthened China’s resolve to develop its own AI ecosystem, leading to increased domestic innovation rather than slowing progress4.
Despite hardware limitations from export restrictions, Chinese firms like Alibaba have developed competitive AI models, demonstrating resilience in model quality5.
The restrictions have prompted Chinese firms to innovate and optimize their models, leading to impressive results despite hardware constraints, with companies achieving cost-effective solutions6.
Huawei has emerged as a key player in developing alternative AI chips to Nvidia’s products, contributing to the push for local chip production as a direct response to export controls1.
This pattern suggests that overly stringent export controls may be counterproductive, potentially accelerating China’s path to technological self-sufficiency rather than containing it7.