Alibaba is trying to make voice the front door to a wider AI workflow. Panda Daily reported on the launch of CosyVoice Studio, which Alibaba describes as China’s first full-stack voice-productivity platform. The product bundles speech recognition, speech synthesis, and real-time voice interaction around Alibaba’s Qwen-Audio family.
China’s humanoid-robot sector has plenty of signals of momentum: packed exhibitions, new product launches, venture activity, and Unitree’s explosive market debut. The harder signal is whether robots can perform useful work with the reliability that customers expect.
Artificial intelligence in China is no longer confined to model launches, industrial parks, or conference stages. It appears in the more ordinary acts of getting home, buying a drink, choosing a product, or asking for medical advice.
China’s AI-chip competition is often described through processors, memory, and export restrictions. Packaging is less visible, yet it determines whether increasingly dense components can move heat away quickly enough to operate reliably. Pandaily reports that Huatian Technology had developed full-process thermal-design and simulation capability for fan-out packaging.
A new China Academy of Information and Communications Technology count shows a national embodied-AI network that has passed the 70-site mark. TechNode reports that the count was current at the end of June, while another 46 facilities were either being built or planned. The significance is not the number alone.
China’s World Robot Conference has become a stage for increasingly ambitious humanoid demonstrations, but one of this year’s debuts drew attention by going smaller. Pandaily reported that Tiangong Omni, a new addition to the Tiangong robot family, appeared at the Beijing event at 1.35 meters tall and 39 kilograms.
Alibaba is moving its Qwen family beyond conversation and into the mechanics of using software. On August 20, Pandaily reported the release of Qwen-UI-Agent, a foundation model for graphical user interfaces that operates across phones, desktop computers, browser environments, and deep-search scenarios.
SenseTime has made a new bid for relevance in China’s crowded visual-AI market by releasing an open model designed to collapse several image tasks into one system. On August 21, TechNode reported that SenseTime made SenseNova U1.5 Lite publicly available. The eight-billion-parameter system brings visual understanding, image creation, and editing together.
Parcel sorting is a punishing test for a robot because the work begins with disorder. Boxes overlap, soft packages sag, labels point in the wrong direction, and the pile changes every time an item is removed.
China’s AI industry is confronting a less visible constraint than model training: the machinery needed to serve models at scale after they have been built. South China Morning Post reports that companies are redesigning software to manage a sharp rise in inference demand while their access to high-end Nvidia processors remains limited.
