Meituan's Trillion-Parameter AI Model Heightens Pressure on Alibaba as Chinese Stocks Trade 15x P/E
BABA•Meituan released LongCat-2.0, a trillion-parameter AI model trained on 50,000 domestic chips, signaling intensified competition for Alibaba's in-app AI assistance features. International equities trade at roughly 15x price-to-earnings versus U.S. tech at 25x, while improving Chinese regulations are attracting foreign capital toward Alibaba shares.
1. Meituan Unveils LongCat-2.0
LongCat-2.0 is a trillion-parameter model trained entirely on a 50,000-chip Chinese cluster, capable of processing inputs up to one million tokens and designed for agentic coding tasks. Early results show it matches or exceeds leading models on coding and agent benchmarks and it will be released as open-source.
2. International Stocks Offer Valuation Edge
International equities currently trade at roughly 15x price-to-earnings compared to 25x for U.S. technology stocks, reflecting valuation disparities that could benefit Chinese names. Concurrently, easing regulatory measures in China are drawing increased foreign capital into local giants like Alibaba.
3. Impact on Alibaba's Valuation and Strategy
Alibaba faces heightened AI competition as Meituan's model demonstrates domestic chip capabilities, potentially pressuring margins and R&D investment. However, lower comparative valuations and renewed capital inflows with regulatory relief could support Alibaba's share performance and attract new investors.
