Back to news
launchTencent2026-10-30

Tencent Cloud unveils integrated multimodal data processing and model training solution, cutting data prep from 15 days to 2

At the big data session of the 2026 Tencent Global Digital Ecosystem Conference on October 30, Tencent Cloud released an integrated multimodal data processing and model training and inference solution for physical AI scenarios such as embodied intelligence and autonomous driving, covering data proce

At the big data session of the 2026 Tencent Global Digital Ecosystem Conference on October 30, Tencent Cloud released an integrated multimodal data processing and model training and inference solution for physical AI scenarios such as embodied intelligence and autonomous driving, covering data processing, training, inference and simulation validation in one flow. It also reported progress on the data agent workbench DataBuddy, which enterprises in finance, retail and education use for data development, governance and analysis. Vice president Huang Shifei said that as agents become a new class of data platform user, how enterprise data is used and what compute loads look like are changing, and platforms must understand the business, support controlled task execution and carry multimodal processing, training and inference efficiently. Because embodied AI and autonomous driving need large volumes of video and sensor data and iterate through training and simulation validation across previously disconnected tools, Tencent Cloud combined the DataBuddy workflow with data lake compute AI DLC to put data management, development orchestration and model tasks on one platform. Xpark and Meson handle multimodal cleaning, quality control, annotation and sample preparation, TCRay provides distributed training and inference, and TCLake unifies management of data and AI assets, with issues found in validation flowing back into data preparation for the next iteration. The company said the solution cuts data preparation from 15 days to 2, improves GPU utilization nearly 200 percent and lowers average daily cost per card by 35 percent, and is already used by companies including Wuwen Zhike, Zhuoyu Technology and Meichi Technology. Tencent Cloud will also launch the TCLake XPhysis platform covering the full physical AI engineering chain.

腾讯云Physical AI多模态数据训推一体DataBuddy具身智能