Verkada, a player in AI-enabled physical security, has announced a collaboration with NVIDIA to support the deployment of physical AI across built environments. As part of this development, NVIDIA is also described as a new investor, following prior investment from Alphabet CapitalG last year.
Using AI, Verkada processes operational data into actionable insights for sectors including education, healthcare, retail, and manufacturing. The stated objective of these applications is to support safety and operational efficiency.
The collaboration is intended to advance Verkada’s intelligent video analytics using NVIDIA’s technology. This includes the use of NVIDIA’s foundational models and data capabilities to improve AI-based video search, such as multimodal embeddings, enhanced vector retrieval methods, and synthetic data generation for training dataset improvement.
Since the collaboration began, Verkada reports a 68% improvement in mean average precision (mAP) for spatial-temporal understanding in its AI-powered search systems. According to the companies, this has contributed to faster and more accurate search performance with increased robustness.
Additionally, Verkada is developing a multi-model search agent architecture and exploring reasoning models aimed at handling complex, unstructured real-world scenarios. Use cases referenced include identifying health and safety incidents in manufacturing environments and detecting shrinkage in retail, with the broader goal of improving contextual understanding in physical environments.
