At its annual Cloud Summit, Alibaba demonstrated its cutting-edge AI capabilities by launching the self-developed AI chip, Zhenwu M890, alongside a 128-card super-node server and the flagship large-scale model, Qwen3.7-Max. These innovations underscored the latest progress in Alibaba’s full-stack system, which integrates “chips, cloud services, models, and inference.” The Zhenwu M890, developed by T-Head, offers triple the performance of its previous iteration. Additionally, Alibaba disclosed its chip development roadmap, with plans to introduce the V900 in the third quarter of 2027 and the J900 in the third quarter of 2028. To date, T-Head has shipped a cumulative total of 560,000 chips and is making preparations for an independent initial public offering (IPO). In third-party blind testing, Qwen3.7-Max outperformed other domestic models, showcasing its ability to execute long-cycle tasks and exceptional results across various benchmark tests. The model will soon be accessible via API through Alibaba Cloud’s BaiLian platform. Propelled by the rising demand for computational power in enterprise-level agent applications and the push for technological autonomy among major corporations, Alibaba’s three-year investment plan of RMB 380 billion is advancing smoothly.