Meituan Develops 1.6 Trillion-Parameter AI Model Using Domestic Chips
Meituan, a major Chinese technology company, has developed a new artificial intelligence model named LongCat-2.0. This model boasts an impressive 1.6 trillion parameters, making it one of the largest in the field. Significantly, the entire training process for LongCat-2.0 was conducted using AI chips manufactured within China. The development highlights how data generated from Meituan's extensive food delivery operations is being leveraged to advance cutting-edge AI research. This initiative underscores China's growing capabilities in developing advanced AI technologies, particularly its reliance on domestically produced hardware for critical training processes.
Meituan's development of a 1.6 trillion-parameter model, LongCat-2.0, trained on domestic AI chips, signifies a strategic move towards technological self-sufficiency in a competitive global landscape. The utilization of food delivery data, a rich source of real-world transactional and logistical information, demonstrates a potent feedback loop where operational scale directly fuels AI advancement. This approach could offer a competitive edge by creating proprietary AI capabilities, potentially leading to enhanced service optimization and new product development. The reliance on domestic chips also reflects broader geopolitical trends and national strategies aimed at building independent technological ecosystems, mitigating supply chain risks, and fostering domestic innovation in critical sectors like artificial intelligence.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.