TIER IV integrated NVIDIA's world foundation model platform, NVIDIA Cosmos, into its Co-MLOps data-sharing and machine learning operations platform to automate data search, edge-case generation, and data augmentation for autonomous driving. The company addresses the long-tail problem, where rare edge cases tied to collision risk are underrepresented in real-world driving data, by using generative AI to fill data gaps through three approaches: search, create, and augment. TIER IV leads the development of Autoware, open-source autonomous driving software, and has conducted demonstrations in more than 39 prefectures and 127 locations across Japan. In 2024, TIER IV launched Co-MLOps, a collaborative framework unifying data sharing and MLOps to develop ADAS/AD systems at scale. The platform operates as an active learning loop where shared data becomes richer as more organizations participate, and the reference AI continuously improves. The loop cycles through stages including autolabeling, data tagging, data gap reporting, data generation, quality validation, similarity ranking, uncertainty estimation, and privacy anonymization. This autonomous improvement cycle finds missing data, generates or collects it, labels and assures quality, and feeds it into training.