The field of ocean dynamics and autonomous underwater systems is experiencing significant growth, driven by innovative approaches to reconstructing subsurface ocean dynamics, developing modular and accessible AUV systems, and creating benchmark environments for underwater embodied agents. Recent research has focused on addressing the challenges posed by incomplete observations, hydrodynamic disturbances, and degraded sensing in turbid waters. A key direction in this field is the integration of multimodal foundation models with embodied intelligence systems, enabling robust operation under communication and computational constraints. Additionally, the development of open-access modular frameworks and comprehensive benchmark environments is advancing the field by providing scalable and cost-effective solutions for AUV systems and facilitating the development of robust embodied AI. Noteworthy papers include: VISION, which introduces a novel reconstruction paradigm for subsurface ocean dynamics, and UnderwaterVLA, which presents a dual-brain Vision-Language-Action architecture for autonomous underwater navigation. OceanGym is also a significant contribution, providing a comprehensive benchmark environment for underwater embodied agents.