Baidu says Chinese buyers want local AI chips due to ‘supply chain’ issues
Baidu has told investors that demand for its in-house Kunlunxin AI chips will keep growing because Chinese buyers face constrained supply of alternatives, reinforcing expectations that Nvidia's products will remain largely locked out of the Chinese market. Speaking on the company's Q2 earnings call, AI Cloud Group executive Dou Shen said Baidu is pressing ahead with plans to spin off and float Kunlunxin, citing rising inferencing demand and China's "significant growth potential" as domestic chip supply stays limited.
The remarks come despite Washington's move to let Nvidia resume China sales, a reversal that has so far yielded no revenue for the chipmaker amid Beijing's own restrictions on such purchases; Nvidia previously said the export ban cost it $10.5 billion over six months. Baidu's AI cloud revenue climbed 50 percent year-on-year to nearly $1.1 billion, with GPU cloud revenue up 283 percent, though these figures trail rivals such as AWS, Google, Microsoft and Alibaba. The company also highlighted over a million autonomous Apollo Go robotaxi rides in Q2 and an 83 percent rise in daily active users of its ERNIE assistant, while overall revenue grew just 4 percent to $3.9 billion.
- Baidu expects strong demand for its own Kunlunxin AI chips
- Chinese buyers face tight supply of alternative chips like Nvidia's
- Baidu's AI cloud revenue grew sharply, but overall growth stayed modest
New here? Start with this
AI chips are processors designed to handle tasks such as training and running artificial intelligence systems. Baidu, a major Chinese technology company, makes its own chips under the Kunlunxin name and also sells computing services through its cloud business.
Nvidia, a US company, is a leading supplier of AI chips. Sales to China have been affected by US export controls and Chinese limits on purchases, while local chipmakers are developing alternatives. This has made access to computing equipment a concern for Chinese businesses building AI services.
The issue matters because chip supply can affect how quickly companies expand those services and how much they rely on domestic suppliers. Baidu’s chip plans sit alongside its wider AI efforts, including its ERNIE assistant and Apollo Go robotaxi service.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
China faces real and significant vulnerabilities in semiconductor supply due to geopolitical tensions and export restrictions that constrain access to advanced foreign chips. Developing indigenous capabilities like Baidu's Kunlunxin is therefore strategically essential for technological sovereignty and security, reducing dangerous dependence on suppliers subject to political restrictions. Building domestic capacity, even if initially less advanced than alternatives, provides crucial independence for long-term technological progress.
The case against
Framing local chip demand primarily as a response to supply constraints rather than technical merit or cost-effectiveness masks protectionist policies. If local chips were genuinely superior or more economical, they would succeed on merit alone without invoking supply constraints. Fragmenting global technology markets through government-influenced procurement reduces overall efficiency and innovation, ultimately slowing progress rather than serving genuine consumer needs.