Alibaba Group Holding Limited (BABA) Earnings

Management highlights

- Strategic Priorities - The company maintains two core strategic priorities: AI + cloud, and consumption. Investments in these areas are already delivering tangible commercial results. - The global AI industry is at an inflection point shifting from conversational chatbots to autonomous AI agents, driving explosive growth in training, inference, and agent orchestration workloads. Alibaba's AI business has exited the initial investment phase and entered large-scale commercialization. - AI and Cloud Commercialization - Annualized AI-related product revenue for Cloud Intelligence Group has surpassed RMB 35.8 billion, maintaining triple-digit growth, and management expects AI revenue to exceed 50% of external cloud revenue within one year, becoming the primary growth engine for the cloud business. - Token consumption on the Bailian Model Studio platform grew substantially quarter-over-quarter, as enterprise customers shifted from simple proof-of-concept tasks to production-scale complex workloads. - T-Head proprietary GPU chips have achieved mass scaled production, with over 60% of compute capacity already serving external customers across internet, financial services, and autonomous driving verticals. This gives Alibaba full supply chain autonomy for AI compute and a structural competitive advantage amid global compute scarcity. - Alibaba Token Hub (ATH) continues to launch new products connecting consumer and enterprise environments, with breakthrough progress in AI-native software and coding agents. The Qwen foundation model continues to iterate on reasoning, coding, and agentic capabilities. - AI Consumer Applications - On May 7, the Qwen AI personal assistant fully integrated Taobao and Tmall commerce service capabilities, and is now deeply embedded across the Alibaba ecosystem (Taobao, Alipay, AMAP, Fliggy), making it China's first all-in-one personal assistant that seamlessly connects daily life, productivity, and learning. - Consumption Business Performance - China e-commerce CMR grew 8% YoY on a like-for-like basis, with improved user experience and merchant operating efficiency. Quick commerce maintained stable market scale while delivering significant unit economics improvement, with higher average order value (AOV) driven by order mix optimization. Quick commerce order volume grew 2.7x YoY, with non-food order volume growing 3x YoY.

Guidance

- AI-related product revenue is expected to exceed 50% of Cloud Intelligence Group's external revenue within one year, and external cloud revenue growth is expected to continue accelerating beyond the current 40% rate in coming quarters. - Model and application services annualized recurring revenue (ARR) (including Model Studio) is expected to surpass RMB 10 billion in the June 2026 quarter and RMB 30 billion by the end of fiscal 2026. - Quick commerce unit economics (UE) is expected to turn positive by the end of fiscal 2027, and the business is expected to achieve overall profitability at its new scale and market share over the long term. - Over the next two years, quick commerce losses will narrow substantially, and AIDC will transition from losses to profitability, driving positive net cash flow growth for the consumption business segment. - Cloud Intelligence Group's gross margin is expected to see significant improvement over the next 1-2 years, with visible improvement starting as early as the next 1-2 quarters. - Over the next 3-5 years, aggregate AI compute infrastructure demand will be 10 times the level Alibaba held in 2022, before the generative AI boom. - 2C commercialization of AI personal assistants is expected to emerge in China within the next 1-2 years, matching the business model development seen in the US.

Segment performance

Consolidated total revenue for the quarter was RMB 243.4 billion, with 11% year-over-year like-for-like growth after excluding disposed Sunnah and InTime businesses. GAAP net income was RMB 23.5 billion, up 96% YoY, driven by higher mark-to-market gains on equity investments and prior-year disposal losses, partially offset by lower adjusted EBITDA. Adjusted EBITDA decreased 84% YoY due to strategic investments in AI, quick commerce, and user experience. Operating cash flow was an inflow of RMB 9.4 billion, while free cash flow was an outflow of RMB 17.3 billion. As of March 31, 2026, net cash was approximately US$38 billion (US$59 billion excluding debt maturing beyond five years). 1. China eCommerce Group: Revenue was RMB 122 billion, up 6% YoY. Customer Management Revenue (CMR) grew 1% YoY (8% YoY like-for-like excluding accounting changes from a new merchant subsidy program). Quick commerce revenue was RMB 20 billion, up 57% YoY. Adjusted Ipta was RMB 24 billion, down 40% YoY due to quick commerce and technology investments; excluding quick commerce losses, adjusted Ipta was stable YoY. 2. AIDC: Revenue grew 6% YoY. Adjusted EBITDA loss narrowed significantly YoY, approaching break-even, driven by logistics and operating efficiency improvements. Unit economics for the AliExpress Choice business improved substantially sequentially. 3. Cloud Intelligence Group: External customer revenue grew 40% YoY (11th consecutive quarter of triple-digit AI-related product growth). AI-related product revenue for the quarter was RMB 9 billion, with an annualized run rate of RMB 36 billion, accounting for 30% of total external cloud revenue. Adjusted EBITDA margin remained stable at 9.1%. 4. All Other Segments: Revenue decreased 21% YoY to RMB 65.5 billion, primarily due to the disposal of Sunnah and InTime, partially offset by growth from Freshippo and AMAP. Adjusted EBITDA was a loss of RMB 21.2 billion, driven by increased investments in foundation models and the consumer-facing Qwen APP.

Risks & headwinds

- Global AI compute infrastructure capacity faces physical bottlenecks over the next 3-5 years: chip, memory, and data center construction production capacity cannot keep up with rapidly growing AI demand, leading to severe server cost inflation — the cost of deploying new servers in 2026 is more than double the cost two years ago. - Domestic Chinese semiconductor production capacity is still limited, constraining the rapid scaling of T-Head proprietary chips that would otherwise drive margin improvements. - Current 2C AI assistant consumer willingness to pay remains low in China, slowing near-term commercialization of consumer-facing AI products. - Aggressive, sustained investment in AI infrastructure over the next two years has resulted in near-term negative free cash flow and pressure on consolidated adjusted EBITDA.

Analyst Q&A

• Q: How much of model and application service ARR comes from Alibaba's proprietary models versus third-party models, and what is the impact of recent per-token price increases on cloud margins? /

  A: Model and application service revenue currently consists primarily of API calls on the Bailian platform, plus AI software subscriptions. Most revenue today comes from Alibaba's proprietary models (including the Qwen series, plus image, voice, and video generation models), though Bailian operates as an open platform that also supports third-party open and closed source models. As AI shifts from chatbots to complex autonomous agents, demand for inference has grown sharply, and customers have high acceptance of moderate price increases; demand outpaces current supply capacity. MaaS inherently carries higher gross margins than traditional IaaS, and ongoing inference technology optimization continues to improve token output per server. Combined with ongoing model capability improvements and gradual price increases, rapid MaaS growth will have a strong positive impact on overall cloud gross margins over the next 1-2 years.

• Q: AI investment has driven strong cloud growth but created near-term pressure on group free cash flow and EBITDA. How should investors evaluate AI ROI, and how does management balance aggressive AI investment against earnings stability? /

  A: Negative free cash flow is almost entirely driven by deliberate, high-conviction AI investment, which management will continue at the same aggressive pace over the next two years to capture this once-in-a-generation strategic window. Core Taobao and Tmall operating cash flow remains stable, and over the next two years quick commerce losses will narrow sharply while AIDC moves to profitability, generating improving cash flow from the consumption segment. Growing AI cloud revenue with improving margins will also generate incremental cash flow to reinvest in infrastructure. Alibaba holds a very strong balance sheet with US$38 billion in net cash and strong financing capacity to support continued investment. Management views AI investment as building two core "factories" (training and inference) that will deliver clear, certain ROI over the next 3-5 years, with zero idle server capacity today.

• Q: What are Alibaba's core advantages in the MaaS space relative to other Chinese AI platforms and startups, when will AI coding see similar high growth in China as in the US, and will low Chinese SaaS willingness to pay limit commercial potential? /

  A: Alibaba invests across a far broader range of model types and verticals than most focused AI startups, which are generally partners rather than direct competitors. The company develops capabilities across coding, image, video, voice, and other verticals to meet diverse enterprise demand for multiple model capabilities. China is already seeing the same rapid AI coding growth seen in the US: since late 2025, token consumption growth for AI coding on Bailian has already exceeded 10x. While Chinese customers have historically had lower willingness to pay for traditional SaaS, that dynamic is changing for large model AI services: as long as token value exceeds token cost, demand will be very strong, and this holds equally for China and the US. Bailian ARR for model services is already over RMB 8 billion and is on track to exceed RMB 10 billion in the June 2026 quarter, matching management's target.

• Q: Global AI adoption is seeing faster ROI on 2B enterprise agent workflows, with slower consumer monetization. How does Alibaba prioritize 2B versus 2C AI investment, and will resources shift from Qwen APP to 2B MaaS if enterprise demand outperforms? /

  A: Fundamentally, AI is a computing paradigm shift focused on helping users (both enterprises and consumers) solve problems, so 2B and 2C are equally core long-term priorities. Currently, 2B has higher immediate willingness to pay because ROI is easier to quantify, so most current infrastructure resources are allocated to 2B. However, 2C AI assistants will also develop clear commercial models in the next 1-2 years in China, following the same path already seen in the US, as models improve and prove their ability to solve everyday consumer problems. Management maintains long-term investment in both 2B and 2C.