Alibaba (NYSE:BABA) unveiled its largest and most capable AI model to date on August 3, and the market noticed immediately, sending its Hong Kong-listed shares up about 7% the same day. The new model, Qwen3.8-Max, arrives as Chinese AI developers race each other on a near-weekly cadence, and it is another data point in the case that Alibaba is becoming an AI infrastructure company that still gets priced like an e-commerce one.
The Bull Case: A Frontier Model With A Cloud Payoff
Qwen3.8-Max is built on 2.4 trillion parameters and uses a mixture-of-experts design that activates only about 95 billion of them at a time, cutting compute costs and response delays. In internal testing, it independently executed a software engineering project over 16 days, and on chatbot arena leaderboards it ranked fifth in Text Arena and second on Vision Arena. Alibaba plans to release the model’s weights for public download next week.
That open-weight approach is the point. Alibaba does not need Qwen itself to generate revenue directly. Every developer who builds on it needs computing power, hosting, storage and other infrastructure, and that demand flows straight into Alibaba Cloud. The timing lines up with what the cloud unit has already been doing: Alibaba’s Cloud Intelligence Group reported total fourth-quarter revenue of RMB 41.63 billion (up 38% YoY) and accelerated external revenue growth to 40% in Q4 FY2026, while AI-related product revenue has posted triple-digit growth for 11 straight quarters.
The Bear Case: A Crowded, Undiscounted Race
Alibaba is not winning this race by a wide margin. Qwen3.8-Max’s 2.4 trillion parameters still trail Moonshot’s Kimi K3, a 2.8 trillion parameter model that rattled global stock markets and Silicon Valley when it launched last month. DeepSeek added more pressure on Friday with its V4-Flash model, priced at just $0.14 per million input tokens, a fraction of what larger models charge and, according to Artificial Analysis, cheaper on a per-task basis than Kimi K3, GPT-5.6 Sol or Claude Fable 5. Giving away frontier models for free only pays off if enough of that traffic converts into paid cloud usage, and that conversion is still a bet, not a result.
Alibaba’s core commerce business adds its own drag. Domestic growth has decelerated as Pinduoduo, JD.com and Douyin compete harder, and a quick-commerce price war raises real questions about near-term profitability. The AI buildout itself is expensive, and Alibaba still needs both to keep spending on models and cloud infrastructure and to stabilize a retail business that generates most of its revenue today.
Alibaba Vs. Baidu
Baidu (NASDAQ:BIDU) is Alibaba’s clearest US-listed rival in this fight, running its own AI Cloud business against Alibaba Cloud and positioning its ERNIE model family directly against Qwen. While Alibaba relies on an open-source model strategy with Qwen to drive developer traffic back to its cloud infrastructure, Baidu takes a more vertically integrated approach. Baidu couples ERNIE with its Qianfan Model-as-a-Service platform, targeting enterprise contracts, sovereign and state-owned entity deployments, and autonomous systems like its Apollo Go fleet. This sets up a dual-front battle in China’s enterprise AI market: Alibaba competing on cloud scale and model accessibility, and Baidu competing on turnkey enterprise integration and targeted domain applications.
Hedge fund ownership of Alibaba fell from 115 to 102 funds last quarter, while Baidu’s fell from 57 to 50, so institutional interest is pulling back slightly for both companies. Alibaba trades at a forward P/E of 19.19 as of August 4, versus Baidu’s 24.27, a cheaper multiple despite Alibaba’s bigger AI and cloud footprint. Short interest is lower on Alibaba too, at 1.79% of float compared with 3.68% for Baidu.
What Does This Mean?
Qwen3.8-Max gives Alibaba another data point for the argument that its cloud business, not its storefronts, will define its next decade, but it lands in a market where Moonshot and DeepSeek are moving just as fast and just as cheap.
While we acknowledge the risk and potential of BABA and BIDU as an investment, our conviction lies in the belief that some AI stocks hold greater promise for delivering higher returns and doing so within a shorter time frame. If you are looking for an AI stock that is more promising than BABA and BIDU and that has 10,000% upside potential, check out our report about this cheapest AI stock.
READ NEXT: 10 Best Future Stocks to Buy Under $10 and 12 Best Performing Semiconductor Stocks to Invest In.
Disclosure: None. Follow Insider Monkey on Google News.
