Verkada and NVIDIA team up to accelerate “physical AI” for buildings

Verkada, a company that sells AI-powered physical security and operations tools, announced a collaboration with NVIDIA to accelerate “physical AI” across the built environment. The effort, disclosed in a July 1 report, is framed as a way to move from collecting video and sensor data to using AI to interpret conditions in buildings and other physical settings.

As described in the announcement, the partnership is intended to speed both the development and the deployment of physical AI, rather than focusing only on research prototypes. Verkada did not, in the published report, lay out specific timelines, performance targets, or the precise technical stack it will use in customer deployments.

NVIDIA is also described as joining the initiative, indicating that the graphics and accelerated-computing vendor will provide underlying compute and AI-enablement. NVIDIA has long positioned its platform strategy around accelerating AI workloads, which can be relevant for companies seeking to run computer-vision models at scale and in real time.

Verkada’s push toward physical AI reflects a broader industry shift in security and operations. Traditional surveillance systems emphasize recording and alerts, while “physical AI” aims to detect, classify, and interpret events such as activity, anomalies, or operational states from visual and other inputs. The value proposition typically centers on reducing manual review and helping operators respond faster, while improving consistency in how incidents are identified.

For NVIDIA, partnerships like this are part of its stated push to expand AI beyond purely digital settings into domains that require heavy perception workloads, such as manufacturing, logistics, healthcare, and retail environments. The company’s public newsroom and product messaging often ties AI acceleration to real-world use cases where inference, model iteration, and deployment efficiency matter.

Still, important details were not disclosed in the report. The announcement did not specify whether Verkada will embed NVIDIA’s software libraries directly into its platform, whether it will use NVIDIA hardware in its own operations, or whether customer deployments will run on specific cloud services or on-premises configurations. It also did not describe any licensing terms, pricing changes, or new contract structure.

Investors and customers will likely watch for follow-on disclosures that clarify how the collaboration affects Verkada’s product roadmap. In particular, future communications could indicate what kinds of physical AI capabilities are prioritized, what deployment environments are supported, and how performance and accuracy are measured as models roll out to customers.