This report presents the results of the inaugural ICDAR 2025 Handwritten Notes Understanding Competition, centered on evidence-based question answering over complex scientific handwritten notes. The final competition test set included over 2,000 real academic note images and 1,000 curated questions across diverse STEM domains, characterized by varied layouts, diagrams, and dense equations. The overall challenge was structured as a single task with a multi-phase evaluation protocol designed to assess not only answer correctness but also grounded reasoning ability, requiring models to localize the exact visual evidence supporting their answers. Hosted on the Robust Reading Challenge (RRC) Portal, the competition ensured fair, private-set benchmarking, with 6 valid submissions out of 16 registered teams. The results underscore the obstacles that current state-of-the-art (SOTA) vision-language models (VLMs) face in multimodal reasoning over handwritten content. This first edition marks a promising step toward interpretable, layout-aware document understanding and sets the foundation for future iterations.