Cainiao officially launched its Supply Chain AI Decision Platform at the 2026 Apsara Conference. The platform adopts a general large model plus specialized small model architecture, has been deployed in automotive, semiconductor and retail industries, and its business has grown over 60% annually. Huang Zhong, general manager of the Supply Chain AI Business Unit, elaborated on the platform's timing, technical approach, industry cases (including a delivered Mixue project) and competitive landscape.
Alibaba Cloud and Kingdee announced at the 2026 Yunqi Conference AI SaaS Forum on September 23, 2026 that they will deepen their AI full-stack partnership, expanding from model services to development efficiency, Agent engineering, data and cloud infrastructure, and ecosystem collaboration. Kingdee's Lingji model runtime has been fully moved to Alibaba Cloud's Qwen AI platform, Qoder has entered Kingdee's R&D workflow on scale, and over 60% of Kingdee's code is generated and committed by AI.
On September 23, the OpenAnolis Community held the SkillHub recommendation system launch and quality skill/award ceremony at the AI Agent Native Operating System sub-forum of the 2026 Apsara Conference, attended by six guests including Ma Tao, Liu Dong, and Raghu Nambiar. The recommendation system evaluates skills from four dimensions and outputs overall and category rankings. Awards were announced for best individual contributor, best practice, and best ecosystem enterprise, with eight companies including Alibaba Cloud, Inspur, Intel, and AMD winning the first batch of Best Ecosystem Enterprise Award.
On September 23, 2026, Qiuzhi Technology launched the DEGO S1 life logger at the Yunqi Conference. The thumb-sized wearable AI camera offers magnetic attachment, first-person recording, an AI agent and voice assistant, plus a crowdsourced data-collection mode to expand real-world data sources for embodied AI.
Wen Zheng announced a comprehensive upgrade of Alibaba Cloud's Qwen AI platform, centered on bringing agents into production, adding agent services and industry AI solutions on top of model services, and introducing API Express mode, Agent Studio and Qwen AI cockpit solutions. He also reported that Alibaba Cloud's MaaS customers grew sixfold in a year, compute scale is expected to exceed 20GW by 2032, model cold-start time dropped from 1,200 to 70 seconds, and first-token latency fell 38%.
On September 22, Alibaba Cloud launched its next-generation full-stack self-developed high-performance storage CPFS at the 2026 Yunqi Conference, targeting next-generation AI training clusters. It provides 100 TB/s-class throughput and 100 million-level IOPS, with single-file-system capacity expanded fivefold to 100 PiB. In actual training, model startup time dropped 50%, peak compute utilization rose 30%, and AI storage costs fell 69%.
At the Yunqi Conference on September 22, 2026, Alibaba Group CEO Wu Yongming said that multimodal models integrating understanding and generation are an important research direction, and that AI needs perception and interaction capabilities so people can communicate with it as naturally as with other humans.
At the Alibaba Cloud Apsara Conference main forum on Sept 22, Honor product line president Fang Fei outlined the company's next-phase view of AI terminals: MagicOS will evolve into Agentic OS, a partner-style multimodal agent operating system, and Honor will co-create five vertical-domain models and solutions with Alibaba based on MagicOS and Qwen Intelligence, including the industry's first system-level Agent Harness architecture deployed in MagicOS 11. The new YOYO achieves 91.8% overall task accuracy and 90% end-to-end task closure; the Honor Magic9 series is scheduled for official release on Sept 28.
At the Apsara Conference on September 22, Alibaba showcased Qwen series progress including Qwen4 in training with a new architecture, Qwen4.5 and Qwen5 in development, unveiled Qwen-Image-3.1, released music model Happy Shrimp 1.1, showed world model HappyOyster 2.0 Preview and omni-modal Qwen3.8-Omni, with the next video generation model set for November.
Zheng Sishou announced at the Apsara Conference that Qwen will accelerate Personal Agent development, using the agentic capabilities of the Qwen 3.8 series models to build a new Personal Agent form and serve 300 million users, reaffirming the goal of giving everyone an AI assistant.