Coram AI raises $35 million to make existing security cameras searchable by AI

Why it matters

Coram AI is applying autonomous-vehicle AI talent to a bigger installed base: existing cameras in schools, warehouses and workplaces. The round shows investors backing software that can make physical infrastructure searchable before robotic security becomes mainstream.

Ashesh Jain and Peter Ondruska's Coram AI has raised $35 million in Series B funding to expand its AI-powered physical security platform, Business Insider reported, bringing Coram AI's total funding to $66 million.

The round was co-led by Ansa Capital and Battery Ventures, with participation from UP.Partners, 8VC and Mosaic Ventures. Coram AI did not disclose its valuation, a useful omission in a market where AI security software is being sold against both legacy camera vendors and high-multiple physical AI companies.

Jain and Ondruska are not coming at cameras as security-industry lifers. Jain, Coram AI's CEO, previously led autonomy at Lyft's self-driving division and worked as an engineering leader at Zoox. Ondruska led AI research at Lyft and Toyota's Woven division after Toyota acquired Lyft's self-driving technology in 2021. Their move from roads to buildings is the core of the Coram AI pitch: use perception and search systems built for messy real-world environments, then apply them to the flood of video that most organizations already collect but rarely use in real time.

"Our earlier mission was to protect every public road, and now it is to protect every physical space," Jain told Business Insider.

The bet is software over installed hardware

Founded in 2022, Coram AI connects to customers' existing cameras, badge readers, visitor logs and emergency systems. That matters because physical security budgets are usually tied up in installed infrastructure. Replacing cameras across schools, warehouses, churches or retail sites is a harder sale than layering search, alerts and workflow automation on top of equipment customers already own.

Jain told Business Insider that more than 1,500 sites across the US and Canada use Coram AI, including Hershey's Ice Cream, Lakepointe Church in Dallas and 1-800-GOT-JUNK?. He also said Coram AI has found traction with school districts, where the system can detect a firearm, automatically alert 911 and help trigger a campus lockdown.

Those customer and growth figures remain company-supplied. Jain told Business Insider that since Coram AI's $13.8 million Series A in 2025, revenue has grown fourfold and its customer base has tripled. Coram AI did not disclose revenue, retention, pricing or gross margin, so the new financing is easier to read as a distribution and product-development round than as proof of category dominance.

Deep Investigation turns footage into a query layer

Coram AI is also rolling out Deep Investigation, a product that lets customers ask natural-language questions across security data and access the tool through systems such as ChatGPT or Claude, according to Business Insider.

Jain described two use cases: a school asking Coram AI to find every fight on campus over the past month, identify patterns, blur student faces and generate a report; or a warehouse asking how long trucks sit at loading docks and when a facility is busiest. The point is not only incident response. Coram AI is trying to turn security cameras into an operations dataset.

That is why the company sits at the intersection of two markets: physical security and physical AI. Business Insider cited PitchBook data showing global robotics and physical AI venture investment rising from about $4 billion in 2019 to $26 billion in 2025, with more than $23 billion raised so far in 2026. Coram AI's version of that trade is less capital-intensive than building robots: sell AI software into buildings before the robots arrive.

Jain's longer-term ambition is to connect Coram AI to robotic security systems, including robot dogs and potentially humanoids. "Every building will have robotic security," Jain told Business Insider. "That is a part of our mission."

Security AI inherits the industry's trust problem

The opportunity is large because most security footage is still reviewed after something happens. The risk is that AI security vendors operate in environments where mistakes, breaches and overcollection can have immediate consequences.

Business Insider named Verkada, Motorola Solutions' Avigilon and Eagle Eye Networks as competitors. The Verkada comparison is especially relevant: Business Insider cited Verkada's $5.8 billion valuation and noted its 2021 breach, when hackers accessed more than 150,000 live customer cameras, including cameras inside hospitals and schools.

Coram AI's product argument is speed: detect incidents earlier, search footage faster and automate parts of response. Its market argument is founder-market fit: Jain and Ondruska spent years building AI for autonomous vehicles, then redirected that background toward security infrastructure already deployed across campuses, warehouses and public-facing facilities.

Investors saw that pitch in live form. UP.Partners investor Ally Warson told Business Insider that during a visit to Coram AI's Sunnyvale office, Jain had Coram AI's software running on office security cameras and invited her to query it in any language. "I told the product in French to pull up all the people who had entered the office that morning wearing white shoes with a blue stripe," Warson said. "And we watched it translate the French into English, and pull up photos of us entering the office."

That demo captures both sides of Coram AI's Series B. If the software works reliably, cameras become searchable infrastructure. If Coram AI sells into schools, workplaces and warehouses at scale, the harder question will not be whether AI can find the clip. It will be who is allowed to ask, what they are allowed to know and how much physical-world data customers are ready to hand to a security platform.