TwelveLabs Raises $100 Million to Make AI Understand Video, Not Just Search It
TwelveLabs raised a $100 million Series B co-led by NEA and NAVER Ventures, with Amazon and Radical Ventures also participating, according to the company's release and reporting from SiliconANGLE and citybiz. The round brings TwelveLabs' total funding to more than $207 million.
The company's original product was video search: letting businesses query large video libraries the way they'd query text, finding a specific moment or scene without manually tagging footage first. The new funding goes toward moving beyond that into a fuller agentic system built on two of its own models, Marengo 3.0, which handles video embedding, and Pegasus 1.5, which converts video into structured data an application can act on. Funds are also earmarked for expanding the company's San Francisco and Seoul teams and opening new offices in New York and London.
A narrower bet than the general-purpose agent race
Most of the venture money flowing into "AI agents" right now is chasing general-purpose capability, agents that can handle arbitrary tasks across text, code, and workflows. TwelveLabs is staying in a narrower lane: video specifically, and the shift from search to structured understanding is really a bet that video is valuable enough as a category, security footage, media libraries, sports and broadcast archives, that a company can build a durable business serving it well rather than trying to be everything to everyone.
That specialization is also what makes the round easier to read clearly than most AI funding news. TwelveLabs isn't claiming to compete with OpenAI or Anthropic on general capability, and its investor list, an oil-adjacent conglomerate arm is notably absent here compared to some of the larger AI infrastructure rounds this year, skews toward strategic investors with a direct interest in video specifically. NAVER, the Korean internet giant, and Amazon both have obvious uses for better video understanding in their own products. When the investors buying in are also plausible customers, it's a reasonable signal the underlying demand is real rather than speculative.
Sources: GlobeNewswire · SiliconANGLE · citybiz