Alibaba (BABA) Earnings Aug 20: Cloud Is 74% of Profit, EBITA Fell 84%

TL;DR

• Alibaba reports June-quarter results before the US open on Thursday, August 20, with the call at 7:30am ET. The company confirmed both.
• Group adjusted EBITA fell 84% last quarter, to RMB5,102m. Cloud EBITA rose 57% to RMB3,796m. Cloud is now 74% of everything the group earns.
• The reported and adjusted profit lines run opposite ways. GAAP diluted EPS per ADS was RMB10.36. Non-GAAP was RMB0.62, down 95%. Alibaba attributes the gap to mark-to-market gains on equity investments.
• Options price a ±5.7% move. BABA has exceeded its implied move in five of the last eight reports, including 10.2% against 5.8% implied in May.
• BABA closed Friday's session at $123.81, up 1.35%, still 35.7% below its 52-week high of $192.67.

The Board

Last quarter's four headline numbers, and why two of them disagree.

When Does Alibaba Report Earnings?

Alibaba reports results for the quarter ended June 30, 2026 before the US market opens on Thursday, August 20, 2026, with a conference call at 7:30am ET. Both come from the company's own announcement.

Worth flagging, because our July piece on the Qwen rally said August 28. That was the date the June quarter landed in 2025. This year the company moved it forward eight days, which puts it in the same session as Walmart and NetEase. I've corrected the older article. The earnings calendar has the rest of the week, and Baidu goes Tuesday.

Where the Headline Profit Came From

Last quarter Alibaba printed net income of RMB25,476m and diluted EPS per ADS of RMB10.36. It also printed non-GAAP diluted EPS per ADS of RMB0.62, down 95% year over year.

Normally the adjusted number is the flattering one. Here it is a fortieth of the reported figure, because the direction of the adjustment reversed: Alibaba's release attributes the net income rise to "the year-over-year increase in net gain from mark-to-market changes of equity investments," plus the absence of the Sun Art and Intime disposal losses that hit the year-ago quarter, "partly offset by the decrease in adjusted EBITA."

So the headline profit came from revaluing a portfolio. The operating business went the other way, and adjusted EBITA fell 84% to RMB5,102m.

One caveat I'll put on the record. The same release gives consolidated non-GAAP net income as RMB86m, against RMB29,847m a year earlier. Divided across roughly 2.28bn ADS, RMB86m is about four fen per ADS, not RMB0.62. Both figures are printed in the filing, so the per-ADS line is evidently struck on a different base. I can't reconcile them from the release, so I'm leaning on adjusted EBITA as the operating measure and flagging the rest.

Cloud Is Nearly All of It Now

Cloud Intelligence Group revenue grew 38% to RMB41,626m, with external customer revenue up 40% and AI products posting triple-digit growth for an eleventh straight quarter. Cloud adjusted EBITA rose 57%, to RMB3,796m.

Set that against the group's RMB5,102m and cloud is 74% of consolidated adjusted EBITA, from a segment that is roughly a sixth of revenue. That ratio is mine, not the company's, and it is the number I'll be watching Thursday.

The other side of it: quick commerce revenue grew 57% to RMB19,988m, and the subsidy war behind that growth is most of what flattened group profit. Management has said the phase of buying share through subsidies is over. The June quarter is the first full period where that claim can be checked, and it matters more than the revenue line.

What the Street Expects, and One Number I Dropped

Consensus revenue sits near RMB268.9bn, about +8.6% on the RMB247.65bn Alibaba posted in the June 2025 quarter. That one divides out correctly, and it sits sensibly between last quarter's RMB243.4bn and the year-ago figure.

A second widely syndicated line put EPS at $1.49 against $2.19 a year earlier and called it a decline of 5.83%. Those three numbers cannot all be true; $1.49 against $2.19 is down about 32%. I couldn't source the pair anywhere primary, so it stays out.

The Options Angle

The market is pricing roughly ±5.7% for Thursday. I think that is light, and BABA's own history is the reason rather than a general view that volatility is cheap.

This stock has exceeded its implied move in five of its last eight reports. May 2026 delivered 10.2% against 5.8% implied. August 2025 delivered 14.3% against 4.6%. February 2025 delivered 14.9% against 7.3%. The misses were genuine duds, moves of 0.4-0.5% against implieds above 6%, so this is a fat-tailed distribution rather than a reliably underpriced one.

What makes a wide move plausible this time: two segments moving hard in opposite directions, a profit line whose adjusted and reported versions disagree by a factor of forty, and the first clean read on whether quick-commerce losses have stopped growing. That is a lot of ways to surprise in either direction, which is the case for owning a straddle rather than picking a side.

I could not source a live option chain, so the structure below is quoted against the implied move rather than a dealt price.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Conviction	Breakeven
1	Long volatility	Long straddle	$124 strike, Aug 21 weekly	~5.7% of spot, no live chain sourced	$123.81 (Aug 14 close)	±5.7%	6/10	needs >5.7%, roughly $131.4 or $116.2
2	Pass	Short premium / covered call into the print	Aug 21	n/a	$123.81 (Aug 14 close)	±5.7%	7/10	n/a

Row 2 carries the higher conviction of the two. Selling a 5.7% implied into a name that has cleared it five times in eight quarters is the exact reflex the August 3 scorecard was written to stop. The week ahead has the rest of the retail and China prints landing around it.

The One-Line Read

Cloud earns three-quarters of Alibaba's operating profit on a sixth of its revenue. Thursday is when we find out whether the rest of the company has stopped burning the difference.