Alibaba's AI Business Is Booming, But Its Profits Basically Disappeared

Alibaba’s fiscal Q4 revenue rose 3% year on year to 243.4 billion yuan, below the 247.1 billion yuan consensus, while cloud revenue jumped 38% to 41.6 billion yuan and AI-related product revenue reached 9 billion yuan. The offsetting weakness was severe: adjusted EBITA for the group fell 84% to 5.1 billion yuan, free cash flow swung to a 17.3 billion yuan outflow, and adjusted net income collapsed to about 86 million yuan from roughly 30 billion yuan a year earlier. Shares fell about 2% premarket as investors weighed rapid AI/cloud growth against heavy subsidy-driven spending and shrinking profitability.

Analysis

Alibaba is in a classic capex-to-monetization transition, but the market is likely underestimating how asymmetric the outcomes are. If AI revenue keeps compounding while cloud margin mix improves, the earnings trough could prove temporary; the key inflection is whether cloud growth remains self-funding before the quick-commerce burn normalizes. The next few quarters matter more than the last print, because once infrastructure is in place, incremental AI load should carry much higher operating leverage than the current P&L suggests.

The real second-order effect is competitive pressure on the broader China internet ecosystem. Alibaba’s willingness to subsidize logistics and AI compute could force peers into either margin defense or strategic retrenchment, especially in lower-tier cities where consumer loyalty is shallow and price elasticity is high. That said, a prolonged subsidy war would also transfer value to logistics providers, fulfillment partners, and hardware/IDC vendors while destroying returns for the platforms themselves.

Consensus is likely anchoring on near-term earnings collapse and missing the option value embedded in AI commercialization. The risk is not that Alibaba lacks demand; it is that management overbuilds capacity faster than monetization can absorb it, which would keep cash flow negative for multiple quarters and compress valuation. The market should treat this as a medium-term execution call, not a one-quarter miss: the stock can rerate sharply if management delivers another quarter of triple-digit AI growth with improving cloud EBITA, but it can de-rate quickly if subsidies remain the dominant growth engine.

Key Decisions for Investors

• Buy BABA on weakness over the next 1-2 weeks only if the stock remains below pre-earnings levels and cloud growth holds above 30% in subsequent commentary; target a 15-25% upside on re-rating if AI monetization stays on track, with a hard stop if free cash flow remains deeply negative for a second quarter.
• Express the view via a call spread in BABA 3-6 months out rather than outright equity; the risk/reward favors upside convexity because sentiment is depressed, but the position should be capped in case capex guidance rises again.
• Pair trade: long BABA / short a China consumer-internet basket with weaker balance sheets or more subsidy exposure; this isolates AI/cloud upside while hedging a broad China demand slowdown and captures relative margin divergence over 2-4 quarters.
• Avoid chasing the move in NVDA on this print; Alibaba’s in-house chip and model strategy is a long-horizon demand signal, but near-term spillover to NVDA revenue is likely incremental rather than material, so the cleaner trade is BABA-specific.
• Set a tactical exit rule: if management signals quick-commerce subsidies will stay elevated beyond the next two quarters, reduce exposure by 30-50%; that would indicate the growth story is still being purchased rather than monetized.