Multimodal artificial intelligence (AI), combining imaging, electrophysiological signals, genomics, and electronic health records, has emerged as a natural response to the heterogeneity of cardiovascular data, yet its clinical translation remains limited by modality misalignment, missing data, and deployment cost. This work reviews current multimodal AI strategies for cardiovascular prediction, from fusion architectures to federated learning and missing-modality handling, and situates this progress within a structured mapping of the European research ecosystem, screening 190 EU-funded cardiovascular projects to identify 83 explicitly involving AI or computational modeling.