Airis Labs exits stealth with $60M to transform video intelligence for defense and government

Defense-AI Startup Raises $60M to Process Unstructured Visual Data

Airis Labs emerged from stealth mode on Tuesday after two and a half years of quiet development, announcing $60 million in total funding including a $31 million Series B funding round led by PSG Equity1

1

. The Tel Aviv-founded defense-AI startup attracted backing from TLV Partners, Stepstone Group, Redseed Ventures, and notable angel investors including Eyal Waldman, the former Mellanox co-founder whose company Nvidia acquired for roughly $7 billion in 20201

1

. PSG director Rotem Shacham, who led the investment, previously served in Unit 8200, the Israeli intelligence unit that also shaped much of Airis Labs' founding team1

1

.

Video-First Intelligence Platform Addresses Critical Government Challenge

The company built what it describes as a video-first intelligence platform designed specifically for government agencies struggling with overwhelming volumes of visual information2

2

. The software ingests fragmented footage from security cameras, drones, body-camera recordings, smartphone uploads, and social media imagery, then produces machine-readable output that analysts and AI agents can query and act upon1

1

. According to Airis Labs, the platform can accelerate some AI video analysis tasks by a factor of 150 while eliminating the need for multiple processing tools2

2

.

Intelligence Teams Drowning in Hours of Unindexed Footage

The problem Airis Labs targets reflects a documented crisis across military and law enforcement organizations. Urban incident investigations routinely generate thousands of hours of mixed-source footage, while drone mission archives fill terabytes with unindexed video1

1

. Human analysts cannot review this volume quickly enough to identify operationally relevant signals before they lose value. The platform addresses this indexing challenge by enabling users to configure monitoring for specific locations, displaying events of interest with natural language descriptions showing what happened, where, and when2

2

.

Multimodal AI Models Process Video, Audio, and Text

The underlying technology relies on multimodal AI models built on transformer architecture, originally designed for text analysis but adapted to handle various data types2

2

. These neural networks convert videos, audio snippets, and text into mathematical representations called embeddings before processing. A built-in investigation tool provides a drag-and-drop interface that lets analysts quickly retrieve specific clips while the AI extracts key details across different content formats2

2

.

Competing Against Palantir and European Defense Players

Airis Labs enters a competitive landscape where Palantir's Project Maven remains the flagship US defense-AI deployment for video analysis, while Helsing operates at the European end with broader battlefield-AI capabilities1

1

. Anduril, Scale AI's defense unit, and Israeli startups including BlueGreen Vision and ION-X compete on overlapping territory. Airis Labs differentiates itself by claiming its platform was built in real operational environments from inception rather than productized from research labs, suggesting tighter user-experience and workflow integration1

1

.

What Government Buyers Should Watch For

Founded in 2023 by Noam Friedman (CEO), Rotem Abeles (chief product officer), and Amos Lahav (chief business officer), the company draws talent from Palantir, Meta, and Israel's Unit 8200 intelligence community1

1

. The startup has been selected into the Oracle Defense Ecosystem and claims operational deployment across government organizations worldwide1

1

. However, Airis Labs has not disclosed which specific agencies are paying customers, revenue figures, or operational benchmark comparisons against named competitors. "Government teams do not have a shortage of raw visual data. They have a shortage of machine-readable understanding," said CEO Friedman2

2

. The company's validation claims rest on analyst feedback rather than published benchmarks, leaving procurement transparency questions for the next 18 months to resolve.