The rapid expansion of single-cell and spatial omics technologies have enabled vast, complex, and multimodal datasets that raise major computational challenges, including processing, integration, and analysis. For many CPU-based approaches, it is increasingly challenging to accommodate the rising needs of modern studies where analysis, integration, and joint processing of millions of cells, fine-scale cellular/tissue maps, and different layers of omics involvement is required. GPU-based computing capabilities hold great promise to address this urgent need with particular capabilities in vast parallelization and reduction of processing time. This special session wants to focus attention for the exploration of recent advances in GPU-enabled methods, specifically designed for single-cell and spatial omics analysis, covering a broad area of scalable pre-processing, quality control, dimensionality reduction, clustering, spatial modeling, classical and modern machine-learning-based inference. It also concerns emerging software ecosystems, reproducible GPU workflows, and deployment strategies for GPU pipelines across local workstations, HPC infrastructures, and cloud environments. The session brings together computational scientists, method developers, and domain researchers with the objective of identifying the existing state-of-the-art solutions, existing bottlenecks, and the direction for the future for scalable and high-performance research analysis. The long-term objective is to lead to the next generation of analysis tools that are native to GPUs and could meet the growing need for single-cell and spatial biology.