Alibaba announces June quarter 2026 results on Thursday, August 20, before the US market open, with the call at 7:30 a.m. Eastern. It reports as a foreign private issuer on Form 6-K, so there is a release and a call and no 8-K.
The consensus card at the top of this page carries the expectations and the two bars this quarter has to clear. We wrote two days ago that Alibaba joined coverage with a Rule of 40 of about one — $145.4B of trailing revenue growing 5.3% against negative $6.6B of free cash flow. The burn is the thesis. Two programmes are pulling on that line at the same time: a multi-year AI datacentre build behind the Qwen model family, and an instant-commerce subsidy war fought on gross margin. Both are deliberate: asked about return on capital on the May call, management said the negative free cash flow follows "very significant investments … in AI over the past year" and that it intends to be "equally resolute" over the next two years, backed by roughly US$38bn of net cash. That is a call remark in a third-party transcript, not a filing.
The June quarter is the first print that can show either of them returning anything. It will not show it in the score — the arithmetic below pins the Rule of 40 into a six-point band whatever Alibaba prints — so the evidence has to come from four specific numbers instead.
The five quarters going in
| Quarter | Revenue | Revenue YoY | Gross margin | Free cash flow | FCF margin | Diluted EPS/ADS | R40 |
|---|---|---|---|---|---|---|---|
| 2025 Q1 (Mar) | $32,584M | +6.0% | 38.4% | $516M | +1.6% | $0.71 | +7.6 |
| 2025 Q2 (Jun) | $34,571M | +3.3% | 44.9% | −$2,626M | −7.6% | $2.51 | −4.3 |
| 2025 Q3 (Sep) | $34,808M | +3.3% | 39.2% | −$3,068M | −8.8% | $1.23 | −5.5 |
| 2025 Q4 (Dec) | $40,732M | +6.1% | 40.5% | $1,622M | +4.0% | $0.85 | +10.1 |
| 2026 Q1 (Mar) | $35,283M | +8.3% | 34.5% | −$2,508M | −7.1% | $1.50 | +1.2 |
Three of the last four quarters are negative on cash, gross margin is at its thinnest of the five, and the Rule of 40 column — quarterly revenue growth plus quarterly free-cash-flow margin — sits below eight in four of the five. Revenue growth here is in dollars and carries the exchange rate with it, which is worth a paragraph of its own further down. The row that matters most on Thursday is the second one: the June 2025 quarter is the one that drops out of the trailing twelve months when the new print lands.
The trailing line improves by $2.6 billion before Alibaba does anything
Trailing free cash flow is a four-quarter sum, and the quarter dropping out of it is the June 2025 quarter, which was negative $2,626M — the first of the negative run. So the arithmetic sets two bars, and neither is a matter of opinion:
- Trailing free cash flow improves for any June-quarter print better than −$2,626M.
- Trailing free cash flow reaches zero at a June-quarter print of +$3,954M.
That second number is the interesting one. Alibaba has not generated $3.95B of free cash flow in a quarter since the December 2024 quarter's +$5,346M, and it has printed negative in three of the last four. Call it the optimistic case rather than the expected one.
Now run it through the score. Growth is the other term, and it is not knowable in advance, so here it is at three rates against the June 2025 base of $34,571M:
| June-quarter FCF | Trailing FCF | R40 at +5% growth | at +8% | at +11% |
|---|---|---|---|---|
| −$3,000M | −$6,954M | +0.9 | +1.7 | +2.5 |
| −$2,626M (flat YoY) | −$6,580M | +1.2 | +2.0 | +2.7 |
| −$1,500M | −$5,454M | +2.0 | +2.7 | +3.5 |
| $0 | −$3,954M | +3.0 | +3.7 | +4.5 |
| +$1,622M (best of last four) | −$2,332M | +4.1 | +4.8 | +5.6 |
| +$3,954M (trailing FCF to zero) | $0 | +5.7 | +6.4 | +7.2 |
The whole grid sits between +0.9 and +7.2. The bottom-left corner is a bad quarter and the top-right corner is a quarter better than any Alibaba has printed in six, and the distance between them is six points of Rule of 40. A company burning cash at this rate for this long cannot be re-rated by one quarter of it, and any read of Thursday that turns on the score is reading the wrong line.
The same holds on Alibaba's own renminbi basis, which is the honest cross-check. Fiscal 2026 revenue of RMB 1,023,670m grew 3% reported and 11% like-for-like, against fiscal-year free cash flow of −RMB 46,609m, an FCF margin of −4.55%. That is a Rule of 40 of −1.6 on the reported growth rate and +6.4 on the like-for-like one. Every basis we can build lands in the same narrow band.
What is actually checkable on Thursday
The score is inert. The commitments are not. On the May 13 call, management put four specific, dated numbers on the record, and the June quarter is the first print that can grade any of them. All four come from third-party transcripts — Alibaba publishes none — so they are attributed claims, not guidance in a filing:
- Model and application services ARR above RMB 10bn in the June quarter. Management said ARR was "already over RMB 8 billion" and that "this quarter, it's highly certain that we can achieve ARR of over RMB 10 billion," with RMB 30bn by year-end. That is a named figure for the exact quarter about to print, with the word "certain" attached.
- Cloud gross margin visibly improving "in the next 1 to 2 quarters." June is the first of the two. Alibaba does not disclose a cloud gross margin, so the checkable proxy is Cloud Intelligence Group's adjusted EBITA margin, which was 9.12% in the March quarter against 8.03% a year earlier — 1.1 points, on our arithmetic from the segment tables in the March-quarter capture.
- AI-related product revenue past 50% of cloud revenue "within a year." It was RMB 8,971m in March, about 30% of Cloud's external revenue and 21.6% of the segment's total, in its eleventh consecutive quarter of triple-digit year-over-year growth. The twelfth is the number to watch: this is the entire reason anyone is funding the capital programme, and a break in the streak is the most consequential thing that can happen in this print.
- Quick-commerce unit economics positive by the end of fiscal 2027, meaning the March 2027 quarter. June is checkpoint one of four. Quick commerce grew 57% to RMB 19,988m in March with order volume at 2.7× the year-ago quarter.
Point three has a comparator worth holding in mind. Baidu's June quarter, which we covered on Monday, carried GPU Cloud growing 283% inside an AI Cloud Infra segment that fell 17% sequentially — a triple-digit AI line inside a segment going the other way. Alibaba's version of that tension is the same shape and much larger.
We will judge the top line on like-for-like, and we are saying so now
Reported March-quarter revenue grew 3%; like-for-like grew 11%. The eight-point wedge is the Sun Art and Intime disposals plus a change that moved platform subsidies from sales-and-marketing expense to contra revenue against customer management revenue. There is a real question about whether that adjustment still applies to the June quarter or whether the disposals have annualised out of the comparison, and we do not know the answer before the release.
So: we will read the top line on the like-for-like basis, stated before the print rather than chosen after it. If the wedge has closed, the two numbers converge and this commitment costs nothing. If it has not, we are on record as to which one we used.
One further caveat on our own figures, and it cuts against the number in our headline. The trailing dollar series here is Alibaba's own US dollar convenience translation, and the renminbi appreciated across the comparison: the March 2026 quarter translated at RMB 6.8980 to the dollar against RMB 7.2568 a year earlier. That is why the March quarter grew 3% in renminbi and 8.3% in dollars. Roughly five points of the trailing 5.3% dollar growth rate in our Rule of 40 is the exchange rate, not the business.
What to watch
- Free cash flow against −$2,626M and +$3,954M. The first is the level at which the trailing line stops deteriorating; the second is the level at which a year of burn nets to zero. Capital expenditure is not a statement line in Alibaba's release — it appears only in the investing-activities narrative — so the free-cash-flow figure is the company's own non-GAAP disclosure and there is no second way to derive it.
- Model and application services ARR. RMB 10bn was called "highly certain" on the last call. It is the only forward number management attached that word to.
- AI-related product revenue, and whether the triple-digit streak reaches twelve quarters. RMB 8,971m in March, ~30% of Cloud's external revenue.
- The "All others" adjusted EBITA line. It swung from a RMB 3.4bn loss to a RMB 21.2bn loss in March — that single line was about 62% of the group's entire year-over-year adjusted-EBITA decline, and it is where the Qwen consumer app's user-acquisition spend lands. Whether it widens tells you more about the AI programme's cost than any capex figure will.
- Operating margin. March came in at −0.35%, the first operating loss in the tracked window, with gross margin at 34.51% against 38.4% and cost of revenue at 65.5% of revenue, up 3.9 points. That is the subsidy war above the line. Note also that GAAP net income nearly doubled in March purely on a RMB 33.8bn mark-to-market swing on investments; non-GAAP net income was RMB 86 million, down 100%. Expect the same divergence to need untangling on Thursday.
What the print is worth, on our own model
We published a forward model for Alibaba alongside this preview — four verticals on the reportable segments, eight quarters of restated history, twenty quarters out. It is worth reading against Thursday for one result, which surprised us.
The base case values Alibaba at $120.66 a share against the $128.90 close, so the model is not arguing the stock is cheap. What it does say is where the argument lives. The Capex peak case — the March quarter was the top of the capital programme, so cloud earns its margin earlier — is worth $3.46 a share against base. The bear case, in which cloud growth decelerates and the subsidy war has no end, is $66 below base. The bull case is $107 above it.
So the question everyone will ask on Thursday, whether the burn has peaked, is worth about three dollars. The questions Thursday cannot answer — how fast cloud compounds for five years, and what multiple a Chinese ADS eventually earns on it — are worth a hundred and seventy between them. At 2.0x terminal revenue the model lands exactly on today's price, which is the honest reading of where the market and the model disagree: not about the cash flows, about the multiple.
What Thursday has to prove, assumption by assumption
The value of publishing the model the day before the print is that every assumption in it becomes a bar the release can clear or miss. Six of them are checkable on Thursday morning. These are the base case's own numbers for the June quarter, written down before it exists:
| Thursday's disclosure | The model input it tests | What the base case needs | What breaks it |
|---|---|---|---|
| Cloud revenue growth | Cloud compounding at 7.0% a quarter, ~31% a year, decaying to 15% | Growth at or near +31% year over year, down from the +38% just printed | A print below ~30% says the decay is already steeper than modelled; another +38% says the model is too conservative |
| Cloud adjusted EBITA margin | Cloud EBITDA margin gliding 20% → 30% | 10.0%, up from 9.12% in March | Flat or lower — and management said the trend would be visible within one to two quarters |
| Group capital expenditure | Cloud capex intensity gliding 35.5% → 22% | Group capex at about 10.9% of revenue, just below March's 11.05% | Anything above 11.05%. The glide is the single assumption that decides most of the fair value |
| AIDC adjusted EBITA | AIDC margin gliding 2.8% → 10% | A positive adjusted EBITA, against −RMB 138m in March | Another loss. This is the most speculative assumption in the model and the least evidenced |
| China E-commerce adjusted EBITA margin | China E-commerce EBITDA margin gliding 22% → 26% | About 20.1%, up from 19.64% | Below 19.6% means the subsidy war is still widening, not stabilising |
| "All others" adjusted EBITA | The folded line's margin gliding −44.8% → −10% | The loss narrowing to about 48% of that line's revenue, from 51% | A wider loss says the Qwen consumer app is still buying users at an increasing rate |
The RMB 10bn ARR figure management called "highly certain" is not a model input — it sits inside cloud revenue — but it is the leading indicator for the first row. Model-as-a-service is the mix management says drives the second row too, so those two lines stand or fall together.
One caveat on reading these bars literally. The model projects smoothly and Alibaba's quarters are seasonal — December is the peak and March the trough — so the margin and capex bars are firmer than the growth bar, which should be read as an annual rate rather than as a forecast of one quarter. And every figure above the segment revenue and adjusted EBITA lines is an assumption of ours, including the depreciation add-back and the split of capital expenditure across segments, because Alibaba publishes neither by segment.
The consensus, and why it is a band rather than a number
Added on the morning of the print, because it is the figure every headline will be measured against and this piece went up without it.
Third-party estimate feeds put revenue at roughly $38.6bn to $39.8bn for the June quarter — about RMB 268bn to 275bn. Against the $34,571M we store for the June 2025 quarter, that is +11.7% to +15.1%, well above the 8.3% the March quarter printed and above anything in the five-quarter table at the top of this page. Note the range carries the exchange rate with it: the same RMB figure converts differently at 6.90 than at 7.15, and Alibaba's own dollar column is a convenience translation at the quarter's rate.
Then there is the earnings line, and it is a mess worth looking at directly. Consensus for non-GAAP earnings per ADS spans roughly $1.49 to $1.94. That is a 45-cent spread — 30% of the low end — on a single quarter of a company that has already reported four of the twelve months in question.
A 30%-wide consensus is not analysts being careless. It is the same divergence this piece flagged three sections ago, priced into the estimates: in the March quarter Alibaba's GAAP net income nearly doubled on a RMB 33.8bn mark-to-market swing on investments, while non-GAAP net income was RMB 86 million, down 100%. When one line can move by that much and the other by that little in the same three months, the earnings number is a statement about which adjustments you accept, and the analysts modelling it do not agree.
So two cautions for this morning, and they are the same caution:
- The stored EPS on our Alibaba page is GAAP diluted per ADS — $1.50 for March, $2.51 for the June 2025 quarter. The $1.49–1.94 above is non-GAAP. Subtracting one from the other produces a number that looks like a result and is not one. We made exactly that mistake available to ourselves on Walmart and are not repeating it here.
- We are not going to print a year-over-year change on the consensus. Doing it properly needs the year-ago non-GAAP figure per ADS, which this repository does not hold, and doing it against our stored GAAP $2.51 would manufacture a 32% decline out of two different measures.
Options markets have been pricing an implied post-earnings move of about 5.6% to 6.6%, below the roughly 7–8% Alibaba has actually moved on recent prints. That is the market saying it expects this quarter to be less eventful than the last few. Given a consensus that cannot agree with itself to within 30%, that is a curious thing for it to be confident about.
What the Street is doing while it waits
Consensus is Strong Buy with an average 12-month target of $189.60, about 47% above the $128.90 close on August 19. Barclays is at $195 and Bernstein at $180; Morgan Stanley and Citi both trimmed targets in July — to $180 and $192 — while keeping Buy. Nobody in that set is rating the cash flow. They are rating cloud, which is the same bet management is making with the capital programme, and Thursday is the first quarterly instalment of the evidence either way.
Sources and provenance. Quarterly revenue, gross margin, free cash flow and diluted EPS per ADS are Alibaba's own US dollar convenience translations from each quarterly release, as tracked on our Alibaba page; free cash flow is the company's reported non-GAAP measure. March-quarter segment revenue, segment adjusted EBITA, cost of revenue, operating margin, non-GAAP net income and the RMB 6.8980 translation rate are from the SEC 6-K Exhibit 99.1 (accession 0001104659-26-060224), reconciled in our March-quarter capture; the RMB 7.2568 prior-year rate is derived from that capture's own RMB and dollar figures. The trailing twelve-month sums, the +$3,954M break-even, the Rule of 40 grid, the Cloud adjusted-EBITA margins of 9.12% and 8.03%, the −4.55% fiscal-2026 FCF margin and the 21.6% AI share of total Cloud revenue are R40 arithmetic on those disclosed figures. The four management commitments — RMB 10bn ARR, cloud gross margin within one to two quarters, AI past 50% of cloud revenue within a year, and quick-commerce unit economics by end of fiscal 2027 — are call-only claims taken from third-party transcripts of the May 13, 2026 call, which we treat as non-authoritative because Alibaba publishes no transcript; none of them appear in a filing. The "equally resolute" and net-cash remarks in the opening carry the same status. The August 20 report date is confirmed by Alibaba IR and carried on our earnings calendar. Analyst ratings and targets are as stored on August 17; the $128.90 price is the August 19 close from our price snapshot and is not what you will see today. The consensus range added on the morning of August 20 — revenue of roughly $38.6bn to $39.8bn, non-GAAP earnings of roughly $1.49 to $1.94 per ADS, and the implied post-earnings move of 5.6% to 6.6% against a 7-8% recent average — is press-reported from third-party estimate feeds. This site stores no consensus series for Alibaba, none of those figures is verified here, and the spread between feeds is quoted rather than narrowed to a point estimate. The +11.7% to +15.1% growth implied by that revenue range is our arithmetic against the $34,571M we store for the June 2025 quarter; no year-over-year change is computed on the earnings line, because the consensus is non-GAAP and our stored per-ADS figures are GAAP.
