Descriptions:
Bloomberg Technology sits down with Citigroup analyst Alicia to unpack why global investors are treating Chinese Internet giants as a “source of funds” rather than as core AI holdings. Despite rapidly improving AI models, many investors are rotating proceeds from Alibaba and Baidu into purer AI plays in Korea, Taiwan, and Japan — frustrated that core businesses like food delivery continue to weigh on earnings even as AI spending ramps up.
On the model front, the conversation highlights meaningful progress from Chinese labs in closing the gap with Western frontier models. Alibaba’s Qwen series (versions 3.5 through 3.7), independent lab MiniMax, and Tencent’s Hunyuan 3 — released in late April 2026 — are all cited as step-change improvements in model intelligence. China’s AI infrastructure spending remains below US hyperscaler levels, partly explained by lower hardware costs and existing capacity built in prior years.
Looking toward the second half of 2026, Citigroup is watching two catalysts: whether Tencent can successfully embed AI agents inside WeChat using proprietary training data, and whether Ant Group’s announced overhaul of the Alipay super-app can bring a functional agentic interface to its massive merchant and mini-program ecosystem. Success on either front, the analyst argues, could trigger a sector re-rating and reverse the fund outflows that have weighed on Chinese Internet stocks throughout the first half of the year.
📺 Source: Bloomberg Technology · Published June 15, 2026
🏷️ Format: Interview
