The partnership marks a significant shift for Verkada as it moves to refine its AI-powered video search, multimodal embeddings, and synthetic data generation. Since integrating NVIDIA Cosmos foundation models and the Physical AI Data Factory, the company reports a 68% improvement in mean average precision for spatial-temporal understanding. These gains allow the platform to better interpret complex, unstructured scenarios, ranging from manufacturing floor safety incidents to retail shrinkage.

Filip Kaliszan, co-founder and CEO of Verkada, noted that the collaboration accelerates the company’s decade-long effort to deploy intelligent systems within the built environment. Beyond immediate search upgrades, the firm is currently developing a multi-model search agent architecture designed to handle reasoning tasks. This transition aims to move beyond simple motion detection toward context-aware systems capable of managing security across hospitals, schools, and global enterprise facilities.