Key Points
- Baidu has announced two new domestically-made AI chips, the M100 and M300.
- The move is a response to U.S. restrictions on the export of advanced AI chips to China.
- Baidu is also launching new “supernode” products that link multiple chips for better performance.
- The goal is to provide Chinese companies with a powerful and low-cost alternative to foreign chips.
The Chinese tech giant Baidu announced two new semiconductors for artificial intelligence on Thursday. The company says these new chips will provide Chinese companies with powerful, low-cost, and domestically controlled computing power.
This move comes as tensions between the United States and China continue to rise, leading to U.S. restrictions on the export of advanced AI chips to Chinese firms. This has pushed many Chinese companies to develop their own processors or find local alternatives.
At its annual Baidu World technology conference, the company said it will launch the M100, a chip focused on AI “inference,” in early 2026. Another chip, the M300, which can handle both AI “training” and “inference,” is scheduled for early 2027. Training is the process of building an AI model by feeding it huge amounts of data, while inference is when the model uses what it has learned to make predictions.
Baidu, which has been developing its own chips since 2011, also announced two new “supernode” products. These products link multiple chips together to boost overall performance, making up for the limitations of individual chips. This is similar to a strategy used by Huawei, another Chinese tech giant.
Baidu’s new supernode products, called Tianchi 256 and Tianchi 512, will be available next year. The company also revealed a new version of its Ernie large language model, which it says is now better at not only processing text but also analyzing images and videos.
