Alibaba reports its September-quarter results before the U.S. market opens on Thursday, with its stock near a two-year low and analysts watching whether AI infrastructure spending can convert into profit. U.S.-listed shares fell 4% on Monday to $90.62, down 38% so far this year, according to Yahoo! Finance.
The tension is structural: cloud revenue is growing fast, group operating income is falling faster, and the market has not yet priced which signal wins.
What analysts expect from BABA's September quarter
Analysts expect GAAP earnings per share to jump 151% from the June quarter to $1.37, on revenue growth of just 2% to $40.49 billion, according to Koyfin data reported by Yahoo! Finance. Adjusted EPS is forecast to rise to $1.65 from $1.25, while EBITDA is expected to climb 5% to $6.03 billion.
Year-over-year estimates point the same direction. Analysts expect EPS of 10.97 yuan, up from 4.59 yuan, on revenue of 269.60 billion yuan, roughly 11% growth, according to Fiscal AI. EBITDA is forecast at 40.81 billion yuan, up from 17.74 billion yuan.
The consensus target implies a 104% upside from current levels. The lowest target, $95.75, still sits 6% above Monday's close. Of 41 analysts, 39 rate the stock Buy or better, with eight at Strong Buy, 31 at Buy, one Hold and one Strong Sell.
Why Michael Burry watches Baidu instead of buying
Michael Burry, the investor known from The Big Short, treats Baidu as an early gauge of AI profitability in China. After Baidu's earnings call on Monday, he highlighted two signals in his Substack chat: more customers paying upfront to reserve GPU cloud capacity, and customers signing longer contracts. BigGo Finance reported that Burry sees rising prepayments and faster payback as the metrics that matter, while noting Baidu's data centers face fewer power constraints in China.
His position is explicit. I have avoided BIDU and will continue to do so, Burry said.
Dou Shen, who leads Baidu's AI cloud business, explained the mechanism on the call.
We are seeing shorter payback periods for our cloud infrastructure investments. With more prepayments and faster payback, we can reinvest sooner and then scale faster.
The gap between Baidu's segments is the data point investors are mapping onto Alibaba. AI cloud infrastructure revenue rose 94% from a year ago, while AI-native marketing revenue fell 15%. Total revenue slipped 1% to 30.8 billion yuan, slightly above the 30.63 billion yuan consensus.
Baidu CFO Henry Haijian He framed the tradeoff for investors.
We will keep investing in AI, including cloud infrastructures, in-house chips, and foundation models. These investments are critical to our long-term competitiveness, but they will limit our near-term profit improvement.
Burry's exit from Alibaba
The same cost logic drove Burry out of Alibaba. I moved my entire Alibaba position into JD not too long ago, he said in August, with plans to move most of the money back within a month or two. After Alibaba announced a $10.2 billion share sale to fund AI investment, Burry said the stock would have to fall by half from its then-current level before he would buy again.
Can Alibaba turn cloud growth into profit?
The June-quarter numbers define the question. AI Cloud and Compute Services revenue rose 45% to 48.44 billion yuan, with an adjusted EBITA margin of about 12%. Group operating income fell 57%, weighed down by tech spending, a goodwill impairment and a provision.
On Monday, Alibaba launched Qwen4-Max, claiming the model cuts average cost per task by about 60% versus its predecessor on the company's workplace benchmark. Efficiency per token, not raw growth, is becoming the metric that determines whether cloud margins scale.
The same day, Washington announced new licensing rules for overseas data centers serving China-based companies, tightening access to computing power from Nvidia chips. The constraint pushes demand toward domestic compute, and Stocktwits traders noted that the restrictions could raise demand for Alibaba's own chip unit, T-Head. Other users on the platform read the curbs as a bearish signal, with sentiment nonetheless rated extremely bullish for both BABA and BIDU on extremely high message volume.
Training data practices have also drawn attention. Shanghai Jiao Tong University researchers reportedly used about 5,000 answers from OpenAI's o1 model to create 10,750 bilingual examples for training an earlier Qwen model. Separately, Alibaba researchers used GPT-4 and GPT-4o to generate multilingual training material, then scored the responses before using them to train smaller Qwen models.
Read the protocol, not the sentiment
Alibaba's setup is a clean test of AI economics: revenue growth at 2%, EPS growth at 151%, cloud up 45%, operating income down 57%. Thursday's results will show whether the compute layer starts paying back on Baidu's timeline, prepayments up, payback periods shortening, or whether the capex cycle extends and margins compress further. Baidu's year-to-date decline of 43%, against Alibaba's 38%, suggests the market has already priced in a delay.
FAQ
When does Alibaba report its September-quarter results?
Alibaba reports before the U.S. market opens on Thursday.
What do analysts expect for BABA EPS?
Analysts expect GAAP EPS of $1.37, up 151% from the June quarter, on revenue of $40.49 billion, according to Koyfin data cited by Yahoo! Finance.
Why is Michael Burry watching Baidu?
Burry treats Baidu's rising GPU prepayments and shorter payback periods as an early gauge of AI infrastructure profitability, though he says he has avoided the stock.
