By Saadet Gokce
ISTANBUL (AA) - Chinese technology company Baidu on Thursday unveiled two new artificial intelligence chips, escalating competition amid the global race to secure advanced computing capabilities.
Baidu introduced the M100, which is aimed at improving inference efficiency in models using the mixture-of-experts technique, and the M300, designed for training super-large multimodal models with trillions of parameters, the South China Morning Post reported.
Shen Dou, Baidu’s executive vice-president and head of its cloud unit, said the M100, set for early 2026, and the M300, planned for 2027, will provide "powerful, low-cost, and controllable AI computing power" to support China’s drive for AI self-sufficiency.
Baidu aims to gain more than 50% performance improvement in serving AI systems, with the Tianchi256 stack scheduled for release in the first half of next year.
An upgraded version, Tianchi512, integrating 512 chips, is expected in the second half of 2026.
By 2030, the company plans to build a supernode capable of supporting "millions" of chips.
Baidu’s founder and CEO, Robin Li Yanhong, said the AI industry is "very unhealthy" due to a system that disproportionately rewards developers of chips and foundational AI models.
Earlier this month, White House spokesperson Karoline Leavitt said the US will not allow Nvidia to sell its most advanced AI chip to China for now.
Trump had also said that the most cutting-edge chips produced by the most valuable corporation in the world would be restricted to US companies and barred from sale to China and other nations.
Li said about half of the new software code Baidu produces is now generated by its AI coding platform, Baidu Comate.
"But 50% is not enough; I hope that in the future we can reach 80 or 90%,” he said.
On Thursday, Baidu announced its latest natively omni-modal foundational model, ERNIE 5.0, on the US social media company X.
The model is capable of omni-modal understanding, creative writing, instruction following, among other functions.