JPMorgan: Alibaba Cloud Profits to Cover AI Losses for the First Time in Q4 2026
nashnova research
JPMorgan expects Alibaba Cloud's EBITA profit to fully offset AI-unit losses by the December 2026 quarter — calling it 'the first key data point for AI capex ROI' — and maintains an Overweight rating despite cutting the price target.
What does "cloud profit covers AI losses" actually mean?
Alibaba's AI labs and applications are currently loss-making. JPMorgan estimates ~RMB 10.6 billion in EBITA losses for the September 2026 quarter alone.
Those losses have been absorbed by cloud-segment profits. This means → as long as cloud earns less than AI burns, the combined segment EBITA stays negative.
JPMorgan forecasts cloud margins crossing 15% and revenue growth topping 55% YoY in the December quarter, pushing combined EBITA into positive territory for the first time.
In plain terms = Alibaba's AI spending would no longer need subsidies from other business lines — the cloud segment digests it internally.
How much more bullish is JPMorgan than the Street?
JPMorgan's FY27 non-GAAP EBITDA forecast is RMB 193.2 billion — 14% above consensus.
For FY28, the forecast is RMB 290.9 billion — 30% above consensus.
This means → the Street has neither fully priced in the December-quarter profit inflection nor factored in 50%+ cloud growth in FY28. The latter is the main driver of JPMorgan's outsized FY28 estimate.
This reflects a view that the market is systematically underestimating Alibaba Cloud's growth curve.
The price target was cut — so why still bullish?
US ADR target cut from $210 to $200; Hong Kong target from HK$205 to HK$195.
FY27/28 total revenue forecasts trimmed by 2% and 6%; non-GAAP EPS forecasts cut by 9% and 10%.
Yet JPMorgan simultaneously raised its valuation multiple from 16x to 17x, citing reduced uncertainty in the earnings path.
In plain terms = near-term earnings estimates are lower, but the *confidence* that those earnings will arrive is higher — so the stock deserves a richer multiple. Overweight rating maintained.
How high can cloud margins go long-term?
JPMorgan sees cloud margins reaching ~20% as infrastructure investment matures.
The report flags a risk: long build cycles mean depreciation hits before utilization ramps, dragging first-year cloud revenue.
This means → depreciation is a real cost, and shareholder returns ultimately depend on whether asset ROI can exceed the cost of capital.
What's happening on the e-commerce side?
Taobao/Tmall EBITA is expected to be flat YoY in the September quarter, an improvement from a 3% YoY decline last quarter. This reflects a marginal easing in competitive intensity.
Instant retail (e.g. Ele.me) losses are estimated at ~RMB 9.2 billion this quarter, down sharply from RMB 35 billion a year ago.
This means → a rising share of higher-value products is lifting average order value, narrowing losses without sacrificing market share.
What's the next thing to watch?
The December 2026 quarter is the make-or-break checkpoint: can cloud + AI combined EBITA actually turn positive on schedule?
If it does, JPMorgan's 14%-30% gap versus consensus gets its first hard data point.
If it doesn't, the logic behind the multiple upgrade weakens.
市场有风险，内容仅供研究参考，不构成投资建议。