Baidu, a Chinese tech company, has released two new AI processors, the M100 as well as the M300, at its annual Baidu World conference. These chips were made by Baidu's Kunlunxin Technology division. The M100 chip is designed for tasks that involve making decisions or predictions (inference) and will be released in early 2026. The M300 chip is built for training large AI models with trillions of data points and will be launched in 2027.
Baidu plans to release a new AI system called the Tianchi256, which will use 256 of its P800 chips, in the first half of 2026. Later in the year, they will launch an even bigger system, the Tianchi512, with 512 chips. Baidu says these new systems will perform more than 50% better than their previous models. By 2030, Baidu aims to expand these systems to use millions of chips.
According to coherent market insights, the Semiconductor Market is expected to grow at a CAGR of 8.6% from 2025 to 2032. Currently, the market is at USD 585.80 Bn in 2025 and is expected to be around USD 1,043.66 Bn by 2032. The market is seeing a fast inclination towards miniaturization as well as improved chip performance, supported by innovations in semiconductor materials and manufacturing processes. Also, the push for energy-efficient and high-speed computing solutions is pushing investments in advanced semiconductor design, while geopolitical factors are supporting regional supply chain diversification to mitigate risks and maintain stability in semiconductor production.
Both aim to deliver "powerful, low-cost, and controllable AI computing power," according to Baidu Cloud president Shen Dou. Baidu has been developing its own processors since 2011 and now joins domestic chipmakers such as Huawei's HiSilicon, Cambricon, MetaX, and Biren in reducing reliance on NVIDIA hardware. Similar to Huawei's CloudMatrix 384 system based on Ascend 910C chips, Baidu is building new "supernode" clusters that combine hundreds of its in-house processors.
Baidu's founder and CEO, Robin Li, said the current AI industry is "unhealthy" because most profits go to chip makers and foundational AI models. He suggested changing this structure so that more value is placed on real-world AI applications. Baidu also introduced Ernie 5.0, a new AI model with 2.4 trillion parameters, which is twice as big as Alibaba's Qwen3-Max and Moonshot's Kimi K2. Baidu says Ernie 5.0 performs better in tasks involving text, images, video, coding, and AI agents, thanks to its improved training.
Source:
Press Release: Baidu
Company: Baidu
