Twelve Labs raises $100M Series B for video superintelligence; Amazon AWS strategic deal

Twelve Labs, a video intelligence startup, closed a $100 million Series B round co-led by NEA and NAVER Ventures, with Amazon as a major investor and participation from Radical Ventures, Korea Investment Partners, Index Ventures, Quadrille Capital, and Red Bull Ventures. The funding brings cumulative capital to over $200 million. Unlike video generation (CapCut, Runway), Twelve Labs focuses on making vast existing video archives searchable and usable: indexing surveillance systems, sports footage, and enterprise libraries that sit mostly unindexed and inaccessible.

The company's Marengo 3.0 video embedding model understands sound, words, and motion across time. Pegasus 1.5 converts video into structured data: scene boundaries, entities, temporal segments. The new funding positions Twelve Labs to expand from enterprise video understanding into a full-stack agentic system that combines perception, knowledge, and reasoning—with the goal of letting machines extract intelligence from the 90% of world data that lives in video.

For architects: Amazon's investment carries strategic weight beyond capital—the company has designated AWS as Twelve Labs' preferred cloud provider, with a multiyear commitment to optimize video inference on AWS Trainium chips and debut new models on AWS first. This is a cloud-lock pattern increasingly seen across AI (Odyssey + Amazon, CoreWeave + Nvidia): investors use capital to secure customer spend, not just equity upside. For enterprises managing large video libraries, Twelve Labs' ability to make footage addressable represents a data-layer unlock comparable to how search engines made text indexable.