Alibaba and Tencent Q1 2026 earnings hit by AI spending

The headline number from Alibaba's March quarter was an adjusted EBITA collapse of 84% year-on-year to just $740 million ^[1]. That figure has been widely framed as the cost of AI capex, but the underlying mechanics are more layered. Alibaba is simultaneously funding two capital-intensive bets: a generative AI infrastructure buildout anchored by the Qwen model family, and Taobao Instant Commerce, its quick-commerce push into same-hour delivery. The combined effect drove sales and marketing expenses to 21.5% of revenue versus 13.3% a year earlier ^[1].

That dual-spend structure is what makes the 84% number so jarring. Adjusted net income fell nearly 100% to $12 million and adjusted EPS came in at $0.09 against a $1.12 consensus ^[1]. Operating cash flow dropped 66% to $1.36 billion and free cash flow turned sharply negative ^[10]. A pure AI capex story would have left non-AI operating leverage relatively intact. The fact that it didn't tells investors the margin compression is structural to Alibaba's current strategic posture, not a one-quarter accounting artifact.

Tencent's top-line miss obscured the most underappreciated number in the quarter. Online advertising revenue jumped 20% year-on-year to RMB 38.2 billion, driven by an AI-powered ad recommendation model ^[3]. That is the cleanest piece of evidence in either earnings print that a Chinese hyperscaler is converting model investment into incremental revenue on an existing P&L line, not just a future option.

The context matters. Tencent's overall growth slowed to 9%, its weakest in six quarters, and net profit of RMB 58.1 billion missed the RMB 61.42 billion estimate ^[3]. Domestic games grew just 6% to RMB 45.4 billion ^[4]. Against that backdrop, a 20% ad jump funded by AI targeting is doing real work to defend the consolidated growth rate. Pair that with WorkBuddy now ranking as the most-used productivity AI agent in China by daily active users ^[4], and Tencent's pitch is increasingly that its AI investments produce monetizable surfaces in advertising and enterprise software, while Alibaba's investments compound through a cloud line item that still carries hyperscaler-style economics.

Tencent's Q1 capex came in at RMB 31.9 billion, up 16% year-on-year, and management guided 2026 AI product investment above RMB 36 billion ^[3][7]. R&D and marketing expenses tied to AI are projected to grow 55% and 91% YoY respectively, pushing operating margin down by an estimated 2.1 percentage points to 36.4% ^[7]. None of that suggests the capex peak is anywhere in sight.

Alibaba's posture is similar in spirit if different in mechanism. CFO Toby Xu pointed to eleven consecutive quarters of triple-digit AI-related product revenue growth as the justification for the spending pace ^[2]. Model Studio customers grew 8x year-on-year and AI-related revenue inside the cloud business is now running around RMB 9 billion per quarter ^[2]. The competitive overlay is intense: Tencent unveiled Hunyuan 3.0 (Hy3) and spent roughly RMB 1 billion promoting its Yuanbao chatbot during Lunar New Year specifically to keep pace with Qwen ^[3]. Neither company can credibly slow down without ceding ground in a duopoly-style model race, which means the margin pressure visible this quarter is likely a floor rather than a ceiling for the year.

The market reaction was its own data point. Alibaba and Tencent shed a combined $66 billion of market value in roughly 24 hours after their prints, with Tencent alone accounting for about $43 billion of the loss ^[6]. That is a striking response given that both companies posted defensible AI traction narratives. The deeper reading is that investors are no longer accepting the open-ended Amazon-style reinvestment framing on faith. As one market commentary put it, the concern is not the spending itself but the lack of visibility on returns ^[6].

Outside coverage reinforced that read. Bloomberg Television framed the print as 'Alibaba Falls as AI Vision Fails to Impress,' while CN Wire's market-cap framing trended on X under an 'AI Vision Falls Flat' banner — both pointing to a single sentiment shift rather than company-specific concerns. Discussion on wallstreetbets split along the same fault line: one camp treated the spend as long-duration Amazon-style reinvestment that will eventually look prescient, while the other framed it in sunk-cost terms. A separate thread surfaced Alibaba's 34% cloud price hikes as evidence that monetization pressure is already being passed to customers rather than absorbed on the income statement. Chinese commentary captured by 36Kr explicitly warned that revenue growth below 17% would disappoint investors while only growth above 20% would re-rate the names ^[9]. That is a narrow band, and it implies the next two earnings cycles are the window in which Alibaba and Tencent need to prove the AI revenue ramp can offset the capex curve.

Strip away the quarter and the prints describe two distinct strategic identities. Alibaba is increasingly an AI infrastructure company that happens to own commerce and quick-commerce assets. Cloud Intelligence revenue of $6.04 billion grew 38% on AI demand ^[2], Qwen App passed 300 million monthly active users in February ^[8], AI-related revenue inside the cloud business is now running around RMB 9 billion per quarter ^[2], and Model Studio customers grew 8x year-on-year ^[1]. The bet is that compute and model distribution become the load-bearing layer of the business.

Tencent is positioning as an AI-enhanced consumer and enterprise platform company. Its WeChat data and distribution moat feeds an ad model that is already producing 20% growth ^[3], WorkBuddy turns that same distribution into productivity-agent reach ^[11], and Hunyuan 3.0 serves more as defensive parity with Qwen than as a standalone product line. The market is being asked to underwrite two very different return profiles: Alibaba's hyperscaler-style flywheel where margins compress before they expand, and Tencent's overlay model where AI shows up first as incremental advertising and agent revenue. Bloomberg's framing on X — that Tencent is 'seizing the initiative as agentic AI fever grips the country' — captures why the second profile is currently easier to value, even when the topline misses.