On May 13, Alibaba Group released its financial results for the fourth quarter and full fiscal year 2026. This earnings report may mark a new phase in Alibaba's AI narrative.
Over the past two years, market discussions on Alibaba's AI focused primarily on investments, models, organizational adjustments, and strategic statements. The focus is now shifting, as AI is entering a cycle of commercial returns.
According to the latest report, Alibaba Cloud's external commercial revenue accelerated year-over-year growth to 40%, with AI-related products accounting for 30% of that external revenue. This quarter, Alibaba Cloud's AI-related product revenue reached 8.971 billion yuan, achieving triple-digit growth for the eleventh consecutive quarter.
More importantly, CEO Wu Yongming provided a clearer forward-looking assessment: "Over the next year, the proportion of AI-related product revenue will exceed 50%, becoming the main engine for Alibaba Cloud's revenue growth. Consequently, Alibaba Cloud's external commercial revenue is expected to continue accelerating from the 40% base over the coming quarters."
During the earnings call, Wu Yongming stated that the Annual Recurring Revenue (ARR) for AI model and application services, including the Bailian MaaS platform, has already exceeded 8 billion yuan. It is highly certain that ARR will surpass 10 billion yuan in the June quarter, and is projected to exceed 30 billion yuan by year-end.
This signifies that the growth driver for Alibaba Cloud is shifting comprehensively from traditional computing and storage to models, computing power, and Agent services. AI is no longer just an incremental factor boosting cloud revenue; it is beginning to serve as the primary engine for the cloud business.
This inflection point is not accidental.
The global AI industry is transitioning from technological breakthroughs to a window of commercial realization. Model capabilities are improving, inference costs are declining, enterprise customers are moving from trials to deployments, and Agentic AI is driving a leap in token consumption scale. AI cloud services are becoming the most significant new revenue category in the cloud computing market, following traditional IaaS.
Alibaba is positioned within this window.
**01 Confirming the Inflection Point: Full-Stack AI Investment Begins to Pay Off**
The change in Alibaba Cloud is first reflected in its growth curve. According to Gartner data, Alibaba Cloud IaaS maintained its leading position in the Chinese market in 2025, with its market share increasing to 32.8%. This quarter, Alibaba Cloud's external commercial revenue grew 40% year-over-year, accelerating from 35% in the previous quarter.
The core variable behind this renewed acceleration in growth is AI. This quarter, Alibaba Cloud's AI-related product revenue reached 8.971 billion yuan, accounting for 30% of its external commercial revenue.
According to management's judgment, within the next year, the share of AI-related product revenue will exceed 50%, becoming the main engine for Alibaba Cloud's revenue growth. Consequently, Alibaba Cloud's external commercial revenue is also expected to continue accelerating from the 40% base over the coming quarters.
This indicates a structural shift in Alibaba Cloud's growth drivers. The popularity of AI Agents like "Lobster" is a microcosm of this change. The core of such applications is enabling AI to perform continuous tasks. They can call tools, access files, operate browsers, connect to external systems, and even execute cross-platform workflows in the background.
The implication for cloud providers is direct. A single Q&A session consumes a limited number of tokens. A continuously running Agent, which repeatedly reasons, calls tools, reads context, and completes multi-step tasks, drives sustained consumption. The more Lobster-like applications there are, the more frequently underlying models are called, and the more token consumption on the MaaS platform amplifies.
Alibaba's opportunity lies in not just providing models; it has the Bailian MaaS platform, cloud infrastructure, and application scenarios like e-commerce, office, and lifestyle services. The growth of Lobster-like applications essentially validates a direction: AI is moving from chat windows into workflows, and the revenue entry point for cloud providers is extending from computing resources to model calls and Agent operations.
The most noteworthy aspect is the revenue from AI models and application services. Management specifically mentioned on the call that revenue from AI models and application services is experiencing exponential growth. This revenue is driven jointly by underlying model services and upper-layer AI-native software. The former corresponds to model calls, training/inference, and the Bailian MaaS platform. The latter corresponds to Agents, enterprise-grade AI applications, and AI-native software.
Over the past three months, the scale of token consumption for model services increased significantly compared to the previous quarter. Enterprise customers are shifting from simple tasks to scaled production and complex tasks. Demand for model and application services on the Bailian MaaS platform is also rising continuously.
"API token demand on the Bailian MaaS platform is growing very rapidly. Compared to November and December of last year, May and June of this year have seen growth of over tenfold," Wu Yongming stated. He added that the recent ARR for AI models and application services has exceeded 8 billion yuan, and it is highly certain it will quickly surpass 10 billion yuan. By year-end, ARR is projected to exceed 30 billion yuan.
This will bring about two layers of change.
First, a change in revenue structure. The increasing share of AI revenue means the cloud business no longer relies solely on traditional IaaS expansion. Model calls, API services, MaaS platforms, and Agent deployments are beginning to contribute higher-frequency, more sustained revenue. Particularly with the emergence of the ARR metric, the market can more clearly track the ongoing commercialization capability of AI products.
Second, a change in business quality. Traditional cloud resources face pricing competition pressure, but the core of MaaS and model services is the output of intelligent capabilities. Once enterprises deploy Agents within their operational systems, model calls become part of daily operations. For cloud providers, this type of demand is closer to recurring revenue and is more suitable for tracking commercialization progress using ARR.
This is precisely the significance of Alibaba's sustained full-stack AI efforts. From Bailian token consumption to AI model and application ARR, to the sustained call demand driven by Lobster-like Agents, Alibaba's AI revenue is beginning to have clearer tracking metrics. AI investment is starting to appear on the income statement.
**02 Global AI Industry Enters Commercial Harvest Period**
Alibaba's AI inflection point is not occurring in isolation; the global AI industry is entering the same phase. Over the past two years, the biggest market skepticism about AI capital expenditure has been the heavy investment and slow returns. GPUs, data centers, networks, power, and in-house chips have continuously pressured free cash flow. Initially, investors focused on whose model was stronger and whose capital expenditure was larger. Now, the question has shifted: When will these investments translate into revenue?
Overseas tech giants have begun providing answers. Microsoft's FY2026 Q3 disclosure showed its AI business annualized revenue run rate exceeding $37 billion, a 123% year-over-year increase, indicating AI has become a separately measurable revenue engine. Microsoft also defines the current phase as the "agentic computing era."
Google's signal is equally direct. Alphabet's Q1 2026 report showed Google Cloud revenue growing 63% year-over-year to approximately $20 billion, with cloud backlog nearly doubling to around $460 billion. Behind this growth is the simultaneous release of demand for enterprise AI solutions and AI infrastructure.
Amazon, through AWS, demonstrates that AI infrastructure is translating into revenue. In Q1 2026, AWS revenue grew 28% year-over-year to $37.6 billion. Amazon also disclosed that the annualized revenue run rate for its chip business, including Graviton, Trainium, and Nitro, has exceeded $20 billion and is still growing at triple-digit rates.
These data points collectively indicate a change: Investments in AI infrastructure are entering a return period.
More importantly, the returns come not only from revenue growth but also from improvements in unit economics. Goldman Sachs, in a recent report, estimated that by 2030, consumer and enterprise agents could push global token consumption to 24 times current capacity. Meanwhile, after a rapid decline, token prices are beginning to stabilize, while underlying computing costs continue to fall. Therefore, the token economics for hyperscale cloud providers and model providers may see a gross margin inflection point within the next 3 to 12 months.
Alibaba's position aligns with the pace of overseas giants. On the analyst call, Wu Yongming also outlined the path for Alibaba Cloud's future margin improvement. He stated that with inference technology optimization, large-scale deployment of in-house AI chips, and the increasing share of MaaS business revenue, Alibaba Cloud's overall profit margin will see significant improvement over the next one to two years, with changes observable in the short term within one to two months.
Therefore, Alibaba's AI inflection point should not be viewed solely within the context of the Chinese cloud market. It should be seen within the coordinate of global tech giants collectively entering the period of AI commercial returns. They are in the same cohort.
**03 Valuation Anchor Reset: Alibaba Being Re-understood**
For Alibaba, the strategic significance of AI + Cloud is not merely about adding another growth curve. It may change how the market prices Alibaba.
In recent years, the capital market viewed Alibaba primarily through the lenses of e-commerce competition, consumption recovery, shareholder returns, and organizational adjustments. The cloud business was once considered an important asset, but due to slowing growth, its valuation weight was never fully realized.
AI has changed this issue.
First, a growth engine shift. As the e-commerce business matures, Alibaba needs a new curve with a higher ceiling. Cloud + AI provides this possibility. If the share of AI-related revenue continues to increase, Alibaba Cloud will no longer be just a supplement to group revenue. It has the opportunity to become one of Alibaba's most important growth sources over the next three to five years. Changing the revenue structure is core. Traditional cloud resources are susceptible to price competition. MaaS and model services rely more on model capabilities, platform ecosystems, and customer workflow integration. Once enterprises integrate AI capabilities into core processes, both call volumes and customer stickiness increase.
Second, there is room for margin improvement. The MaaS business model charges based on tokens and call volumes. As scale increases, models optimize, and inference costs decline, marginal costs are expected to decrease. This is also the value of full-stack capability. If only focusing on applications, underlying costs are constrained by upstream players. If only focusing on computing power, revenue easily falls into resource price competition. Alibaba's simultaneous work on chips, cloud, models, platforms, and applications aims to gain stronger control over both the cost and revenue sides. T-Head addresses underlying computing efficiency. Alibaba Cloud handles infrastructure and customer reach. Tongyi Qianwen addresses model capabilities. Bailian addresses enterprise development entry. Qianwen App and enterprise agents address application consumption.
Once this system is operational, what Alibaba sells is not just cloud resources; it sells AI capability.
Third, the group's valuation logic may be reshaped. If the market continues to view Alibaba as an e-commerce company, the valuation anchor is influenced by consumption cycles and platform competition. However, if Cloud + AI revenue continues to accelerate, and metrics like AI revenue share, AI model and application ARR, token consumption, and MaaS revenue are consistently delivered, Alibaba will be re-understood. It will resemble more of an AI infrastructure platform. This is the most noteworthy aspect of this earnings report—it provides a connection point from strategic narrative to financial validation.
