Investors, governments eye Israeli AI startup sifting oceans of raw video for intel

Utilized in wake of Oct. 7 attack, Airis Labs says its platform can scan scads of visual data from smartphones, drones, CCTV, social media and more to find useful info faster than humans

The Hamas-led assault on southern Israel on October 7, 2023, was documented in real time from nearly every angle — by cameras the killers wore themselves, by security devices, CCTV, dashcams, phones, and drones overhead. As teams worked to pull relevant intelligence, including hopes of finding the hundreds missing or abducted, Israel’s security services confronted a labyrinth: the answers were almost certainly in there, if anyone could find them.

Drowning in footage, officials turned to local startups in the private sector developing artificial intelligence systems to help analyze and quickly extract relevant intelligence from the mountain of information.

One of those startups was Airis Labs, a startup specializing in AI-powered visual threat analysis.

“In the evening of October 7, several organizations approached us asking for help to analyze videos coming from a wide range of visual sensors from Gaza,” Airis Labs co-founder Noam Friedman told The Times of Israel. “We had just established the startup in early 2023 around lessons that we saw emerging from the war in Ukraine, which exposed that the world of intelligence was rapidly transitioning from text to video, while existing intelligence platforms weren’t keeping pace.”

“We saw analysts drowning in data to extract the needle from the haystack while critical signals and intelligence were being missed,” said Friedman.

Since April 2023, the startup’s founders — Friedman; Rotem Abeles, a former Palantir Technologies strategist; and Amos Lahav – have been operating under the radar to develop an idea into an AI platform designed to connect and analyze large volumes of raw visual data from multiple sources and different formats and turn them into actionable intelligence for national security and law enforcement agencies.

“We focus on turning videos from different sources into intelligence for operational insights that national security analysts can act on at the speed needed to prevent the next catastrophic terror attack,” said Friedman. “We did not have the luxury of spending years inside the lab building the technology.”

“We developed the AI platform in a real-time operational environment during a war and under real pressure and urgency of a warzone rather than based on synthetic data or generic models, which is what sets us apart from competitors,” Friedman added.

Friedman said that the platform can narrow down and extract relevant information using AI models within minutes, instead of days or months.

Israel’s Defense Ministry declined to comment on the deployment of the startup’s AI intelligence platform when contacted by The Times of Israel.

Airis Labs has already been deployed by several government organizations worldwide, it said. Friedman noted that the startup participated last year in a US Army accelerator program tailored to integrate new technologies, while also working with law enforcement and national security and intelligence organizations.

The company emerged from stealth on Wednesday, announcing it has secured a total of $60 million in funding to date from investors, including PSQ Equity, TLV Partners, Stepstone Group, and Redseed Ventures. Among its angel investors are Mellanox Technologies founder Eyal Waldman, whose daughter was murdered by Hamas terrorists at the Supernova music festival near the Gaza border on October 7, as well as Jeff Horing, co-founder of New York-based venture capital firm Insight Partners.

Airis Labs said it will use the funds to hire more staff and expand its US operations. The startup, which employs 50 employees, mostly at its R&D center in Tel Aviv, also has an office in Washington, DC.

“Airis Labs is the rare company that was born from a deep understanding of the problem,” said Waldman. “The founders understand the mission because they’ve lived it. That’s not something you can replicate.”