Alibaba Cloud is executing a Google-style vertical integration strategy—from proprietary chips to foundation models—that analysts believe could fundamentally reshape its revenue structure by 2031, transforming the business from a cloud infrastructure vendor into a model-as-a-service (MaaS) powerhouse.

Analysts at Citi Research, including Alicia Yap and Nelson Cheung, published a 60-page report on May 11 positioning Alibaba Cloud as "China's Google" and one of the country's strongest AI investment proxies. Citi projects AI-related revenue will surge from RMB 24 billion (US$3.3 billion) in FY2026 to RMB 585.5 billion (US$81.3 billion) by FY2031—a 90% compound annual growth rate that would raise AI's share of total cloud revenue from 15% to 70%.

The fastest-growing segment is expected to be model-as-a-service (MaaS). Citi forecasts MaaS revenue expanding at a 235% CAGR, rising from roughly RMB 1 billion (US$139 million) in FY2026 to RMB 438.6 billion (US$60.9 billion) by FY2031, when it would account for 53% of Alibaba Cloud's total revenue. The shift represents a strategic transition from selling compute capacity to monetizing intelligence itself.

Full-Stack AI Strategy Mirrors Google's TPU-to-Gemini Model

Alibaba's ecosystem now spans four of the five layers in Jensen Huang's "AI value stack"—semiconductors, infrastructure, platforms, and models—excluding only energy generation. That vertical integration forms the basis of Citi's "China's Google" thesis.

Chip Layer

T-Head, Alibaba's semiconductor subsidiary founded in 2018, has reportedly shipped 470,000 units of its Zhenwu 810E AI chip, which some Chinese reports compare to NVIDIA's H20. More than 60% of shipments went to over 400 external enterprise customers, including State Grid Corporation of China and Xpeng Motors.

March 2026 product launches included the RISC-V-based XuanTie C950 CPU for agentic AI workloads and the Panmai 920 400G smart NIC, which has entered mass production for Alibaba data centers.

Infrastructure Layer

Alibaba Cloud operates 94 availability zones across 29 regions globally, including 13 domestic regions and 15 international regions. Additional expansions are planned in Brazil, France, and the Netherlands during 2026.

According to Gartner data cited in the report, Alibaba Cloud holds a 33% share of China's IaaS public cloud market, ranking first domestically and fourth globally with 7.7% market share.

Platform and Model Layer

Alibaba's Model Studio (Bailian) integrates the full Qwen model family alongside more than 200 third-party models, including offerings from DeepSeek, Moonshot AI's Kimi, and MiniMax.

The platform has attracted over one million enterprise and individual users, who have collectively created more than 800,000 AI agents. Management disclosed during the FY3Q26 earnings call that token consumption on the MaaS platform increased sixfold over the previous three months, making MaaS the Cloud Intelligence Group's largest revenue category.

Token Economics Are Reshaping China's AI Industry

China's daily AI token invocations exceeded 140 trillion as of March 2026—a 1,400x increase from roughly 100 billion in early 2024, according to data from the National Data Bureau cited by Sina Finance. China also surpassed the United States in weekly model token consumption for the first time in February 2026, becoming the world's largest token market.

Enterprise adoption is accelerating in parallel. Data from Frost & Sullivan and the China Academy of Information and Communications Technology showed enterprise token consumption reached 10.2 trillion daily in H1 2025, up 363% from H2 2024.

Alibaba's Tongyi Qwen captured a 17.7% share among China's most-used enterprise general-purpose models, ahead of ByteDance's Doubao (14.1%) and DeepSeek (10.3%).

Strong demand also enabled Alibaba Cloud to raise AI compute product pricing by 5% to 34% on April 18, 2026—a move Citi interprets as evidence of improving pricing power amid persistent supply constraints.

Alibaba Commits $114B to AI Infrastructure Expansion

Citi projects Alibaba will maintain annual capital expenditures between RMB 130 billion and RMB 140 billion (US$18-19 billion) from FY2026 through FY2031, totaling RMB 822.9 billion (US$114 billion) over six years.

That investment builds on Alibaba's previously disclosed three-year RMB 380 billion AI infrastructure plan announced in February 2025, which CEO Eddie Wu described as a "once-in-a-generation opportunity."

Wu later stated at the 2025 Apsara Conference that Alibaba intends to expand global data center capacity tenfold by 2032 to support progression "from AGI to ASI."

Management also indicated during the FY2Q26 earnings call that supply-chain bottlenecks could accelerate deployment timelines, as AI demand is expected to exceed compute supply for at least the next two to three years.

Citi forecasts capex as a percentage of cloud revenue will decline from 83% in FY2026 to 17% by FY2031, implying expanding operating leverage as scale improves.

Qwen's Open-Source Ecosystem Becomes a Commercial Moat

The Qwen (Tongyi Qianwen) family serves as the commercial foundation of Alibaba's MaaS strategy. Since launching its first open-source model in August 2023, Qwen has expanded into text, reasoning, coding, vision-language, audio, video, and mathematics domains.

Cumulative downloads on Hugging Face exceeded 600 million by the end of 2025, while the ecosystem now includes more than 170,000 derivative models—making Qwen one of the world's most widely adopted open-source AI frameworks.

Qwen3.6-Plus, released on April 2, 2026, reportedly processed more than one trillion tokens in a single day on OpenRouter, setting a platform record. Artificial Analysis ranked Qwen3.6 Max Preview sixth globally on its Intelligence Index with a score of 52, narrowing the gap with OpenAI's GPT-5.5 (60).

Reuters also reported on May 10, 2026 that Alibaba is deeply integrating Qwen into Taobao, enabling conversational AI shopping across more than four billion product listings on Taobao and Tmall. Qwen is also embedded across Alipay and Amap, giving Alibaba proprietary distribution channels competitors may struggle to replicate.

Competitive Landscape: Infrastructure Leader, MaaS Challenger

Although Alibaba Cloud remains China's overall cloud leader, IDC data shows ByteDance's Volcano Engine controls 49.2% of China's public cloud LLM service market by invocation volume, compared with Alibaba's 27% and Baidu AI Cloud's 17%.

Citi argues Alibaba's long-term advantage lies less in any single segment and more in vertical integration. Proprietary chips can reduce infrastructure costs, internal consumer ecosystems like Taobao and Alipay provide data and adoption channels, and Qwen's open-source ecosystem strengthens developer lock-in.

Eddie Wu has argued that only five or six hyperscale cloud platforms globally will ultimately be capable of delivering mass-scale AI compute services. Citi maintains that Alibaba Cloud's positioning as a "world-leading full-stack AI service provider" places it in a strong position to capture that opportunity.

Revenue Mix Is Shifting from Compute to Intelligence

Citi forecasts Alibaba Cloud's total revenue expanding at a 39% CAGR, rising from RMB 159 billion (US$22.1 billion) in FY2026 to RMB 833.4 billion (US$115.7 billion) by FY2031.

External customer revenue is projected to grow at a 44% CAGR to RMB 666.8 billion (US$92.6 billion), approaching management's stated five-year US$100 billion target.

Within AI revenue categories:

• MaaS (235% CAGR) reaches RMB 438.6 billion
• AI-IaaS (46% CAGR) reaches RMB 128.7 billion
• AI-PaaS (40% CAGR) reaches RMB 18.2 billion

Meanwhile, the non-AI cloud business—despite growing at a respectable 13% CAGR to RMB 247.9 billion—shrinks from 85% of revenue to just 30% of the mix.

The implication is clear: Alibaba Cloud is evolving from a traditional infrastructure provider into an intelligence platform monetizing AI services at significantly higher margins.

Citi reiterated its Buy rating with a US$205 ADR target and HK$204 Hong Kong-listed target, designating Alibaba as its top China AI investment pick.

Related Coverage：

Alibaba Stock Hits Four-Year High as Goldman Sachs Sees AI Infrastructure Revival

Alibaba's Chip Strategy: Defensive Positioning or Commercial Breakthrough?