Alibaba Group Holding Limited (BABA) Earnings

Management highlights

- AI and Cloud Commercialization Acceleration - Alibaba Cloud external revenue growth hit 45% YoY, a 22-quarter high, with broad-based growth across compute, storage, model-as-a-service (MaaS), and AI applications - Annual revenue run rate of AI-related products surpassed RMB 49.5 billion, accounting for 35% of external cloud revenue, with 12 consecutive quarters of triple-digit growth - MaaS ARR exceeded RMB 16 billion as of August 2026, with continued strong growth in customer contract pipelines - Management proactively scaled back low-margin cloud business to improve growth quality, and upgraded Alibaba Cloud to an agentic cloud to capture surging AI agent demand - Full-Stack AI Capability Progress - T-Head (Alibaba's proprietary chip division) has a full portfolio spanning GPUs, CPUs, and networking chips; the next-generation Zhenwu M890 AI processor-powered SuperNode instance launched at commercial scale, serving over 650 customers as of early August 2026, with supply ramping up in H2 2026 - Alibaba cut hyperscale AI data center delivery time to 100 days, a world-leading pace for infrastructure expansion - Alibaba recently released major iterations of its large language, image, audio, video, and music models, all ranked top-tier globally; open-sourced QN 3.8 Max (2.4 trillion parameters) model weights, with the QN model series downloaded over 3 billion times globally and over 300,000 derivative models built on it - AI Application Ecosystem Development - Launched QNWork, an enterprise AI productivity product for workforce scenarios, expected to become a new ARR growth engine - The consumer QN app continues growing its user base and expanding value-added offerings, with 250 million users having tested AI-driven shopping experiences via the app's agentic features - A coordinated commercial flywheel across compute, models, tokens, applications, and monetization is operating efficiently - E-commerce Business Performance and Strategic Alignment - Quick commerce scale grew 45% YoY, with losses narrowing substantially and unit economics improving QoQ; Art Express achieved operating profit this quarter - Completed realignment of e-commerce into four core segments: China e-commerce, quick commerce, international e-commerce, and global B2B - For China e-commerce, focused on strengthening partnerships with leading brands, sourcing high-quality goods directly from industrial clusters, and deploying AI to improve consumer experiences (multimodal search, virtual try-ons) and merchant operational efficiency (data analytics, marketing, customer service)

Guidance

- AI and cloud: Revenue growth is expected to continue accelerating in coming quarters, with EBITDA margin expanding steadily QoQ. Management is highly confident in achieving the long-term target of $100 billion in external cloud revenue by 2030, with a 20% gross margin target. - MaaS business: Confident of hitting the year-end ARR target of RMB 30 billion. - AI labs and applications: The segment loss is expected to continue narrowing in coming quarters as model training and QN app marketing efficiency improve. - Quick commerce: Non-food category transaction volume is expected to surpass food category volume in the next fiscal year, with overall profitability achieved in FY29. Long-term, quick commerce is expected to contribute 30% of total platform GMV as e-commerce's second growth curve. - AI CapEx: The current three-year RMB 380 billion investment plan remains on track, with 190 billion spent as of end-June 2026. AI-related CapEx is expected to reach break-even in ~3 years (shortening to 2.5 years or less as gross margins improve and proprietary chip adoption rises), with positive free cash flow generated for at least two years post break-even. While pursuing 40%+ growth, the business will maintain positive cash flow once the payback period shortens. - Overall: AI is confirmed as Alibaba's most certain long-term growth engine, with sustained disciplined investment planned to drive secular growth.

Segment performance

Alibaba implemented new segment reporting starting this quarter, with the following results: 1. Alibaba e-commerce group: Total revenue of RMB 205.9 billion, up 4% year-over-year (YoY). China quick commerce revenue was RMB 53.3 billion, up 45% YoY. Adjusted EBITDA remained stable YoY at RMB 39.7 billion. 2. AI Cloud and compute services: Total revenue and external customer revenue both grew 45% YoY, a 22-quarter high growth rate. AI-related product revenue hit RMB 12.4 billion in the quarter, with an annual revenue run rate of RMB 49.5 billion, marking 12 consecutive quarters of triple-digit growth. AI-related products account for 35% of external cloud revenue. Adjusted EBITDA margin expanded to 12% sequentially. 3. AI labs and applications: Adjusted EBITDA loss of RMB 13.9 billion, with the loss narrowing quarter-over-quarter (QoQ) due to reduced marketing spend for the QN app. The loss stemmed from increased AI capability investments and higher inference costs for the QN app. 4. All other: Revenue remained stable at RMB 28.8 billion, with an adjusted EBITDA loss of RMB 3.3 billion driven by increased technology investments. Group-wide total revenue grew 9% YoY to RMB 269 billion, total adjusted EBITDA decreased 30% YoY to RMB 27.3 billion, GAAP net income was RMB 10.4 billion (down 75% YoY), operating cash flow increased 11% YoY to RMB 22.9 billion, and CapEx totaled RMB 67.7 billion reflecting heavy AI infrastructure investments.

Risks & headwinds

- AI compute supply is currently constrained industry-wide, with the shortage not expected to resolve until at least 2030, which could limit near-term growth if supply expansion cannot keep up with surging customer demand. - International e-commerce faces near-term headwinds from tariff policy changes and geopolitical uncertainty that pressure growth. - Domestic China e-commerce faces short-term macroeconomic challenges that impact consumer spending. - Long-term AI industry value distribution is uncertain: value may shift across layers (chips, cloud, models, applications) as the industry matures, creating uncertainty for revenue concentration. - Semiconductor component prices have risen, increasing CapEx costs and pressuring near-term margins. - Large model competition is intensifying with growing numbers of open-source models, though management notes this dynamic benefits full-stack cloud providers like Alibaba.

Analyst Q&A

• Q: What drove the sharp quarter-over-quarter increase in CapEx, how will CapEx trend going forward, what is the status of the existing 380 billion RMB three-year budget, and what is the expected return on AI CapEx? /

  A: The full 380 billion RMB three-year AI investment plan remains on track, with 190 billion RMB spent as of end-June 2026. This quarter's higher CapEx primarily reflects uneven hardware delivery cycles and pre-procurement of CPUs to meet surging AI agent demand, plus rising semiconductor component prices — this quarter's figure is not a linear trend for future quarters. AI is an asset-heavy business model requiring upfront CapEx to capture future growth. Current AI CapEx breaks even in 3 years, and can be shortened to 2.5 years or even 2 years via three levers: higher-margin MaaS business and optimized product mix, substitution of expensive third-party chips with cheaper, higher-performing proprietary T-Head chips, and ongoing commercial model innovation. Post break-even, assets generate positive cash flow for multiple years, so we can maintain 40%+ growth while staying cash flow positive as the payback period shortens.

• Q: What are the strategic priorities for e-commerce following the recent business realignment? /

  A: After realignment, we focus on four core e-commerce segments. For China e-commerce, we will strengthen leading brand partnerships, source high-quality goods directly from industrial clusters, and deploy AI to improve consumer experiences and merchant operational efficiency. For quick commerce, we will accelerate non-food category growth and front warehouse expansion, targeting overall profitability in FY29 and 30% of total platform GMV long-term. For international e-commerce, despite near-term geopolitical and tariff headwinds, the business has improved profitability while maintaining growth, with strong long-term potential and rapidly growing local platforms in Turkey, the Middle East, and improving efficiency in Southeast Asia. For global B2B, AI is fundamentally reshaping business models, and our new AI agent tool AxioWork already gained over 50,000 paying merchants shortly after launch, creating new long-term commercial opportunities.

• Q: What are the short-term and long-term growth drivers for Alibaba Cloud, and what is the outlook for hitting the $100 billion five-year external revenue target? /

  A: Short-term, growth will continue accelerating, with 45% YoY growth this quarter marking nine straight quarters of acceleration. The core short-term driver is surging demand for AI inference services, which has turned compute from a cost center into a core revenue-generating asset. Supply constraints allow healthy pricing, and our top-tier models give us strong pricing power for new contracts and renewals, driving margin expansion. Long-term, the core driver is the powerful network and scale effect of the AI cloud platform, which acts as the core infrastructure for all AI workloads, attracting more customers as capabilities expand. Our full-stack investments across chips, infrastructure, and models position us to capture value regardless of how industry value shifts across layers. We are highly confident of hitting $100 billion in external revenue by 2030.

• Q: Is the 30 billion RMB year-end MaaS ARR target still on track, how do proprietary vs third-party models split revenue, and will growing model competition hurt gross margins? /

  A: As of August, MaaS ARR already surpassed 16 billion RMB, and given current strong growth momentum and upcoming new model launches, we remain confident of hitting the 30 billion RMB year-end target. Proprietary models account for the majority of current MaaS revenue, but third-party open-source models also generate meaningful revenue. Gross margins are very similar for both model types. Growing competition and more open-source models actually benefit full-stack cloud platforms like Alibaba Cloud, because customers typically use multiple models from different sources in their AI applications, increasing demand for our cloud infrastructure and inference services.