Large model–driven edge-cloud collaboration for next-gen animation systems

Recent advances in large-scale artificial intelligence models are gradually reshaping the design and operation of modern animation systems. By combining the computational capabilities of cloud platforms with the low-latency advantages of edge environments, edge–cloud collaboration provides a practical foundation for integrating large models into animation pipelines. Building on this background, the paper surveys emerging frameworks that incorporate large models into edge–cloud collaborative environments, with particular attention to distributed inference, federated learning, and adaptive data processing mechanisms. These frameworks aim to balance real-time responsiveness with large-scale computation while supporting scalable and collaborative animation workflows. The review further analyzes key challenges associated with this paradigm, including computational overhead, privacy and security concerns, and the preservation of artistic control. Finally, potential research directions are discussed, covering multimodal integration, human-centered animation systems, and energy-efficient computing. Overall, this survey seeks to clarify the current landscape of large model–driven edge–cloud animation systems and to highlight promising directions for future research and development.