BABA Reverses Off the Miss as Wu Pegs AI at Half of Cloud Revenue Inside a Year

The Reversal

BABA was a negative-print stock for about an hour. The fiscal Q4 numbers landed below consensus on both lines — revenue at RMB 243.38B against a roughly $36.35B Street bar, adjusted EBITA collapsing 84% YoY to RMB 5.1B — and the ADR opened down 3.05% on the day. Then the call started.

By the close the stock had reversed to +6.78% at $143.92, and the HIP-3 BABA perp on Hyperliquid had ripped 9.89% over a four-hour window to $143.80. The perp move is wider than the ADR's daily print because it captured the full premarket-low-to-post-call-high arc inside one rolling window. That's the kind of intraday repricing thin stock perps tend to overshoot — the move is real, but the perp number is the wrapper, not the truth.

Wu's AI Math

The market wasn't reacting to the income statement. It was reacting to Eddie Wu putting hard numbers on a thesis that has so far been mostly narrative.

The headline quote: Wu told the call more than half of Alibaba Cloud's revenue will come from AI inside a year. AI was roughly 30% of cloud revenue this quarter, so 'more than half within a year' is an explicit acceleration call from the CEO, not a 'long-term opportunity' slide.

Wu paired it with an ARR roadmap: AI model and application services annualized recurring revenue above RMB 10 billion in the June quarter and RMB 30 billion by year-end, inside the previously stated five-year goal of $100 billion in combined external cloud and AI revenue. He also flagged that compute demand will push spending beyond the previously announced three-year RMB 380B capex plan, with some of that landing as opex via rented compute. Translation: more spend is coming, and management is choosing to pre-commit traders to a revenue curve to absorb it. The bid showed up because the curve is now numbered.

The Print Beneath the Reversal

The reason this is a reversal and not a clean breakout is that the underlying quarter was genuinely weak.

• Adjusted EBITA: RMB 5.1B, down 84% YoY.
• Group revenue: RMB 243.38B, +3% YoY, light of consensus.
• China e-commerce: revenue +6% to RMB 122.22B but segment EBITA -40% to RMB 24.01B as the company funneled profit into Qwen user acquisition and instant-delivery competition with Meituan and JD.
• Quick commerce: revenue +57% YoY to RMB 19.99B, still loss-making.
• The bright line: Cloud Intelligence Group at RMB 41.63B, +38% YoY, with AI-related product revenue at triple-digit YoY growth for the eleventh consecutive quarter and cloud segment EBITA expanding 57%.

Bloomberg Intelligence's Catherine Lim framed it bluntly: Alibaba effectively redeployed more than 90% of its March-quarter China e-commerce profit into Qwen user acquisition. The bull case is that the cloud and AI segment is large enough and growing fast enough to swallow that redirection. The bear case is that the e-commerce moat keeps eroding while the AI ARR ramp is still a forward promise.

What's Now in Play

Two things changed at the level of the trade.

First, the perp has a fresh anchor. Wu's '>50% of cloud from AI within a year' line and the RMB 10B / 30B ARR steps give the next two quarters explicit checkpoints. If June-quarter AI ARR prints above RMB 10B, the call is being underwritten; if it slips, the entire reversal-day move is at risk because the print itself didn't justify it.

Second, the BABA HIP-3 market reasserted itself as a venue for trading earnings-window volatility on a stock that doesn't have a crypto-native correlate. With about $9.35M in 24h perp volume and 10x leverage available, the book is thin enough that the perp will keep overshooting the ADR on both sides of catalysts like this. The structural risk to long exposure isn't the thesis — it's that incremental Chinese tech headlines (Tencent, JD, regulatory tape) and capex revisions can move BABA hard between earnings, and the perp will amplify each one.