TwelveLabs raised $100M and is opening a London office — what it means for AI Builders in the UK and India

What you need to know

• $100M Series B, announced 1 July 2026, co-led by NEA and NAVER Ventures, with Amazon, Radical Ventures, Korea Investment Partners, Index Ventures, Quadrille and Red Bull Ventures joining. Total raised to date is roughly $150M.
• The team roughly tripled in a year — from about 58 people to about 178 — and TwelveLabs is opening New York and London offices on top of its San Francisco and Seoul bases.
• The product is video-understanding foundation models — the Marengo family for embeddings and search, the Pegasus family for generative video-language understanding. The company frames the mission as building "video superintelligence".
• The builder-facing takeaway: a funded team opening a London office is a hiring signal for the whole UK and European market, and it is a template for how you should read every raise. Fresh capital becomes headcount, and headcount becomes searches for people like you.

The round, in plain numbers

TwelveLabs, headquartered across San Francisco and Seoul, confirmed the round in a company blog post and a wire release on 1 July 2026. Here is the shape of the raise, and how it sits against the company's history.

Two details are worth pausing on. First, Amazon is a repeat investor here, and the round came with a commercial string attached — per the funding announcement, AWS becomes the company's preferred cloud, with new models tuned for AWS Trainium landing there first. Strategic money like that usually signals a scaling phase, not an experimental one. Second, the investor list is unusually cross-border: American venture capital (NEA), Korean strategic and financial backers (NAVER Ventures, Korea Investment Partners), European names (Index Ventures, Quadrille) and even Red Bull Ventures. A cap table that spans three continents tends to hire across three continents too.

You can read the primary sources yourself — the company post at twelvelabs.io and the wire release on GlobeNewswire. Reading the source rather than the coverage is a habit worth keeping; the headline number is rarely the most useful figure in the announcement.

What "video superintelligence" actually means

The phrase is the company's own framing, and it is worth treating in quotes rather than at face value. What TwelveLabs builds today is more concrete and more interesting than the slogan. Its models turn raw video into something a machine can reason over. The Marengo family produces video embeddings — the numerical representations that make footage searchable by meaning rather than by filename or manual tag. The Pegasus family handles generative understanding: summarising a clip, answering questions about what happens in it, extracting scenes, entities and temporal segments as structured data.

The company's stated direction is to move from these standalone models towards a single agentic system that combines perception, knowledge and reasoning over video. Whether that adds up to anything like "superintelligence" is a marketing question, not an engineering one. The engineering reality — video as a first-class, queryable data type — is a genuinely underbuilt corner of the field, and that is precisely why a $100M cheque showed up. If you have shipped anything touching video pipelines, multimodal retrieval, or agent tooling over media, this is a company whose problem space overlaps yours.

Why a single raise is a template, not a one-off

The reason this story sits on AI Tech Connect and not just on a funding tracker is that TwelveLabs is a worked example of a pattern that repeats across the market every week. In the first quarter of 2026, a majority of startup capital went to AI-native teams — we covered the numbers in how 57% of Q1 startup capital went to funded AI teams. On the UK side, AI startups pulled in £8.2bn of venture funding in the first half of the year, which we broke down in the UK AI funding surge of H1 2026. Every one of those rounds turns into the same thing TwelveLabs just did: new offices, new headcount, new searches for people who can build.

That is the mechanism. Capital lands, a company commits to a hiring plan in the press release, and within a quarter or two the recruiters and founding engineers start looking for talent. The question that decides whether they find you is brutally simple: when someone searches for an AI builder in London or Bengaluru with your exact stack, does anything come up?

The office announcements are the part worth reading closely. A company opening in New York and London is not making a real-estate decision; it is declaring where it intends to hire. For a builder in the UK, a Seoul-and-San-Francisco company planting a London flag means roles that were previously an ocean away are about to have a local hiring manager, a local salary band and a visa-free path. For a builder in India, the read is slightly different but no less useful: cross-border cap tables like this one — American, Korean and European money on a single term sheet — tend to build distributed teams, and distributed teams hire remotely from exactly the places that a London or New York office cannot cover on cost. Either way, the funded team is about to widen its net, and the people who surface in that search are the ones who already exist as a findable, specific profile rather than a CV sitting in a drafts folder.

What a Builder should actually do this week

Reading a funding announcement and feeling a flicker of ambition is easy. Converting it into being discoverable is the work. Here is the sequence that actually moves the needle, in order of effort.

1. Make yourself searchable first. Before you polish a CV or draft a cold email, make sure that a recruiter searching by skill and location lands on a page that is unmistakably you. A profile that lists your projects, your stack and your city is doing that work around the clock, whether or not TwelveLabs specifically is hiring today.
2. Vet the company before the company vets you. A fresh raise is exciting, but a tripling headcount is also a warning that culture and process are being rebuilt on the fly. Run the checks in our checklist for vetting an AI startup before you join — runway, cap-table quality, and whether the role is a real need or a hype hire.
3. Show video- or multimodal-adjacent work if you have it. This particular company cares about video embeddings, retrieval and agentic tooling over media. If you have shipped anything in that neighbourhood, that is the project to feature first — specificity is what makes a profile findable for a niche role.
4. Be early where the search happens. Directories reward the people who are already listed when a recruiter starts looking, not the ones who scramble to sign up after a role goes live. Being present before the demand arrives is the whole game.

The bottom line

TwelveLabs raising $100M to chase "video superintelligence" is a good story on its own terms — a cross-border cap table, a serious product in an underbuilt corner of AI, and a clear scaling plan. But the more useful reading for anyone building in the UK or India is structural. This is what a funded AI team looks like from the outside: money in, offices announced, headcount tripling, and a hiring engine spinning up. The teams doing this will find their next builders by searching. The only variable you control is whether you are there to be found.