ModelsThe Decoder· Jun 30, 2026
Meituan Trains 1.6 Trillion Parameter Model on Chinese Chips
Meituan developed LongCat-2.0, a 1.6 trillion parameter AI model trained entirely on domestic Chinese hardware. The project demonstrates that large-scale model training is possible without Nvidia components. This milestone highlights growing self-sufficiency in China's AI infrastructure.
Key points
- →LongCat-2.0 reaches 1.6 trillion parameters
- →Training performed exclusively on Chinese chips
- →Reduces dependence on Nvidia hardware for massive models
Mentioned
MeituanLongCat-2.0Nvidia
Related stories
Taiwan Raids Super Micro Offices in Nvidia Chip Smuggling ProbeThe Decoder · Policy & Regulation→Base44 Launches Proprietary AI Model Amid Defensibility EffortsTechCrunch · Models→Hugging Face Adds All Eval Results to Model PagesHugging Face · Models→DiScoFormer: Unified Transformer for Density and Score Across DistributionsHugging Face · Models→Zhipu AI Releases GLM-5.2 Open-Weight Model Matching Mythos in CybersecurityThe Verge · Models→Wall Street Sees Micron as Next AI Stock After NvidiaTechCrunch · Business→
This is an original summary by Dhanasvi's agents based on The Decoder's public feed. For the complete article, visit the original source. Trademarks and article copyright belong to their owners.