Gasgoo Munich- Moore Threads recently signed memorandums of understanding with 16 partners—including Tsinghua University’s AIR Innovation Center, the Shanghai Jiao Tong University Wuxi Photonic Chip Research Institute, the Huazhong University of Science and Technology Wuxi Research Institute, Jijia Technology, and Guanglun Intelligence—to officially launch the Moore Threads (Wuxi) Industrial Embodied AI Innovation Center. The initiative aims to accelerate the industrial application of embodied AI through collective ecosystem strength.
Image Credit: Moore Threads
Specifically, the newly launched center will use Moore Threads’ KUAE large-scale intelligent computing cluster as its computing foundation. Leveraging full-stack embodied AI capabilities, it will focus on Wuxi’s industrial landscape to drive key technology R&D, commercialization, scenario implementation, and talent development.
In Moore Threads’ view, breakthroughs in embodied AI require integrated cloud-edge-end collaboration: training and simulation on large-scale intelligent computing clusters, followed by edge-side deployment via high-efficiency chips.
Acting on this insight, Moore Threads has bridged the entire chain of "large model training—simulation—edge deployment" in the embodied AI space. Powered by its full-featured GPUs, the company has built a proprietary, end-to-end software and hardware stack. This not only includes the successful deployment of the KUAE large-scale intelligent computing cluster but also complete cloud-edge-end synergy capabilities, offering a one-stop, secure, and reliable domestic computing solution for embodied AI.
Consider synthetic data for embodied AI. Built on its proprietary MUSA architecture, a single Moore Threads chip can simultaneously handle AI computing, graphics rendering, physics simulation, scientific computing, and ultra-HD video encoding and decoding. This breakthrough provides integrated, full-link computing support for the production of synthetic embodied AI data.
The KUAE 1,000-card cluster, driven by the flagship MTT S5000 AI training and inference card, leverages full-precision general computing to deliver stable, efficient power for massive synthetic data production. It supports efficient generalization across dimensions such as pose, material properties, perspective, and environmental conditions for single tasks—shifting the paradigm from limited data collection to large-scale generation.
Recently, Moore Threads formed a strategic partnership with Guanglun Intelligence. The two will leverage Moore Threads’ full-featured GPUs and the KUAE intelligent computing cluster, combined with Guanglun’s proprietary "Solve-Measure-Generate" simulation platform, to jointly develop high-confidence simulation data synthesis solutions.
