On May 20, Alibaba Cloud made an announcement about the successful completion of a comprehensive full-stack Agent upgrade, covering the entire spectrum from 'chip to cloud to model to inference.' As part of this upgrade, the company unveiled a super-node server, which is powered by its in-house AI chip, the Zhenwu M890. Additionally, Alibaba Cloud introduced a new AI product portal named Qianwen Cloud and showcased its flagship large model, Qwen3.7-Max. This strategic move is designed to meet the escalating concurrent demands generated by Agents. By overhauling its infrastructure in a holistic manner, Alibaba Cloud aims to establish China's largest AI factory. The Zhenwu M890 chip stands out with its performance, delivering three times the capability of its forerunner and supporting networking for up to 128 AI chips. In blind testing scenarios, Qwen3.7-Max shone brightly, proving its prowess in autonomously accomplishing intricate engineering tasks. Moreover, the Bailian inference platform has integrated reinforcement learning mechanisms, all the while preserving an open ecosystem.