Alibaba Cloud lifts MaaS revenue with deeper focus on AI coding
Alibaba Cloud has seen a significant surge in its Model-as-a-Service (MaaS) revenue, with a 15-fold increase in token usage during the first five months of 2026, reaching a nine-figure RMB monthly sum. This growth is primarily attributed to the rising adoption of AI agents, which are driving increased token consumption through agentic coding. The company recently launched Qwen 3.7-Max, its latest flagship model, emphasizing enhanced coding performance and long-horizon task capabilities to compete in the rapidly evolving AI landscape. This strategic shift reflects a broader industry trend where major cloud providers, including Google, are focusing on cloud-based agentic AI as a central competitive front.
Alibaba Cloud's strategic pivot towards AI agents and agentic coding highlights a critical shift in the Asian AI market. The dramatic increase in MaaS revenue underscores the commercial viability of AI agents, which consume significantly more tokens than traditional chat interactions. This trend is not unique to Alibaba, as competitors like ByteDance are also vying for market share, albeit with different focuses—ByteDance in video generation models and Alibaba in large language models for developers.
The emphasis on agentic coding and long-horizon task capabilities in models like Qwen 3.7-Max signals a new frontier in AI competition. The ability for AI agents to autonomously manage complex tasks, from coding to resource activation, fundamentally alters how cloud services are consumed and developed. This also necessitates deeper integration across chips, infrastructure, and models, transforming the large model race into a systems engineering challenge where efficiency and value-driven token consumption are paramount. For Asia's tech ecosystem, this means accelerated innovation in AI infrastructure and a redefinition of developer workflows, potentially leading to significant productivity gains and new business models.
