Physical security company Verkada Inc. has taken an investment from Nvidia Corp. and signed a technical partnership with the chipmaker, the two said today, in a deal meant to speed up the artificial intelligence running across Verkada’s 2.4 million connected devices.

The size of the investment was not disclosed but comes seven months after Verkada raised funding at a $5.8 billion valuation in a round led by CapitalG, the growth arm of Google owner Alphabet Inc.

Verkada offers security cameras, door access, alarms, environmental sensors and intercoms, all run through one cloud dashboard. The offering has found strong success, with the company’s technology deployed at 30,000 organizations in 170 countries, including more than 100 of them in the Fortune 500.

Two Nvidia products sit at the center of the deal. One is the Cosmos family of world foundation models. The other is the Physical AI Data Factory toolkit, which is used in part to spin up synthetic footage that fills gaps in training data. Verkada is putting both to work on the models behind video search, the job of finding a specific person, object, or moment across thousands of hours of recordings.

That is where the early payoff sits. Verkada says accuracy on one benchmark, mean average precision for spatial-temporal queries, has climbed 68% since the work began.

“Verkada has been building and deploying physical AI before the term existed,” Filip Kaliszan, co-founder and chief executive of Verkada, said in the announcement. “Working with Nvidia supercharges what we’ve spent nearly a decade building: AI that keeps students safe in schools, protects workers on factory floors, helps retailers prevent theft and enables organizations to operate more efficiently.”

Verkada is building a search agent that draws on several models at once. It is also testing reasoning models for the harder calls, the kind that mean reading a scene rather than matching a label. The company points to two: spotting a safety incident on a factory floor and catching theft in a store.

Nvidia has spent the past two years courting exactly this kind of customer. It has pushed the Cosmos models and a stack of developer tools at robotics, self-driving and factory-floor buyers and backed a run of startups building on its chips. Verkada, sitting on years of camera data, is a fit.

The money also caps a strong stretch for the company, which crossed $1 billion in annualized bookings around the time of the CapitalG round.

That growth comes with exposure, however. In 2021 hackers breached live feeds from about 150,000 of the company’s cameras, including those in hospitals, schools and workplaces.

Image: Verkada