Alibaba Cloud has achieved a more than fivefold increase in daily token revenue compared to early April, according to sources familiar with the matter obtained by Caijing magazine. Monthly revenue has now reached the hundreds of millions of yuan level.

According to Caijing magazine's report published on May 14, Alibaba Cloud set token revenue as a core operational metric in 2026. In early April, the company established a short-term token revenue growth target requiring daily token revenue to increase fivefold by May 15 compared to early April levels. This ambitious two-month goal has now been achieved.

Alibaba Cloud declined to comment on these figures when approached by Caijing. However, on May 13, Alibaba Group released results for its fiscal year 2026 (April 1, 2025 to March 31, 2026), showing Alibaba Cloud revenue reached RMB 158.1 billion yuan (US$21.8 billion), up 28.6% year-over-year. In the fourth quarter of fiscal 2026, cloud revenue grew 38% to RMB 41.63 billion yuan, with external revenue (excluding internal Alibaba businesses like Taobao and Tmall) growing over 40%. AI-related product revenue reached RMB 8.97 billion yuan, exceeding 30% of external revenue for the first time.

During the post-earnings conference call, Alibaba management disclosed that annual recurring revenue (ARR) for models and applications, including the Bailian platform, exceeded RMB 8 billion yuan in fiscal Q4 2026, could surpass RMB 10 billion yuan in fiscal Q1 2027, and is projected to exceed RMB 30 billion yuan by fiscal year-end 2027.

The surge in token consumption stems primarily from three sources: the explosion of AI agents driving increased API calls, commercialization progress of Model-as-a-Service (MaaS) platforms, and computing power consumption from internet and large model companies. The emergence of Agent tools in early 2026 fundamentally changed token consumption patterns, as a single agent task can consume tens of thousands or even hundreds of thousands of tokens, compared to just thousands for traditional AI dialogue tools.

Companies such as Moonshot AI, Zhipu, and MiniMax, which use Alibaba Cloud services, have experienced significant growth. A source at Moonshot AI confirmed to Caijing that the company's revenue in the first 20 days of February alone exceeded its entire 2025 revenue. Zhipu CEO Zhang Peng revealed in late March 2026 that the company's model API ARR reached RMB 1.7 billion yuan in March 2026, a 60-fold year-over-year increase.

Token revenue has become a critical metric across the industry. China's National Data Administration reported that as of March 2026, China's daily token call volume exceeded 140 trillion, representing a more than 1,000-fold increase from 100 billion in early 2024. This marks a shift from traditional cloud revenue metrics focused on computing, storage, networking, and database services.

The strategic importance of token revenue stems from Alibaba Group's broader "Cloud+AI" growth objectives. In March, Alibaba CEO Wu Yongming announced a five-year target for "Cloud+AI" external revenue to exceed US$100 billion annually. In March, Alibaba established the Alibaba Token Hub (ATH) business group, comprising five divisions focused on creating, delivering, and applying tokens.

Global cloud providers are similarly prioritizing token revenue. Amazon Web Services' MaaS platform Bedrock generated several billion dollars in revenue by late 2025, with management projecting it could eventually match EC2's contribution, which exceeds US$40 billion annually. Google reported in its fiscal Q1 2026 earnings that its first-party models process over 16 billion tokens per minute, up from 10 billion in the previous quarter.

However, industry observers note that while token consumption is a critical supply-side metric for cloud providers, it may not fully reflect business value for enterprise customers, who are more concerned with computing costs and predictable pricing than token volumes.

Related Coverage：

Alibaba Cloud Sets 80% Target for China's AI Cloud Incremental Market by 2026

Alibaba Cloud Partners With Nvidia to Accelerate Physical AI Applications