Microsoft and Alibaba implement strategic shifts in AI resource and enterprise deployment

Microsoft is implementing internal controls to manage the high costs of generative AI. Leaked data reveals significant disparities in AI resource consumption among employees, with one individual in the Customer and Partner Solutions department reportedly spending $28,000 in tokens over a 28-day period. To optimize operational costs and combat “compute inflation,” Microsoft is shifting its default internal workloads to OpenAI’s GPT-5.6 Sol model and discouraging engineers from using third-party tools like Anthropic’s Claude for coding tasks.

In a separate development within the AI sector, Alibaba’s Qianwen Office has reached 30 million users within one month of its launch. More than half of these users are enterprise clients, signaling a shift toward embedding AI agents directly into corporate workflows. By integrating with platforms like DingTalk, the service aims to transform individual productivity into organizational assets through automated tasks involving documents, schedules, and cross-system data retrieval.