Baidu Announces M100 and M300 AI Chips to Reduce NVIDIA Dependence
Baidu plans to launch the M100 chip for inference in 2026 and the M300 for training and inference in 2027. The initiative comes after Chinese regulators recommended local companies reduce NVIDIA chip usage. Analysts note hyperscalers seek proprietary chips to avoid supply issues and reduce costs.
Key Highlights
- M100: Inference chip launching 2026
- M300: Training and inference chip launching 2027
- Driver: Chinese regulators recommend reducing NVIDIA use
- Goal: Avoid supply problems and reduce costs
- Market: Generative AI model support
Chip Independence
Baidu's chip development reflects broader Chinese tech industry efforts to reduce dependence on foreign semiconductors. As export restrictions tighten, domestic chip alternatives become strategically essential for Chinese AI development.
Chinese hyperscalers are increasingly developing proprietary chips to ensure supply chain security and cost efficiency.