Higgsfield raises $400M, quadrupling its valuation to $5.4B in eight months
The AI video app maker raised a $400 million Series B led by DST Global at a $5.4 billion valuation, up from $1.3 billion eight months earlier. Higgsfield reports $700 million in annualised revenue and 30 million users.
Higgsfield sells finished workflows — cinema and marketing studios — rather than raw model access, and it is reporting $700 million a year doing it. The money in AI video is moving to the interface layer, which is also where creative control is won or lost.
The striking number in Higgsfield’s round is not the $5.4 billion valuation but the revenue underneath it. A company that sells workflows rather than raw model access, and whose underlying models TechCrunch’s report does not even mention, says it is running at $700 million a year. That puts a price on something the industry has been arguing about for two years: whether the value in AI video sits in the model or in the product people actually use to make things.
The details are straightforward. DST Global led the $400 million Series B, with Goldman Sachs Alternatives, Valor Capital and Tribe Capital also investing. The previous valuation, in January, was $1.3 billion. Founded in 2023 by former Snap executive Alex Mashrabov, Higgsfield reports 30 million users across 200 countries and says it works with 390 of the Fortune 500. Its products are organised by job rather than by model: Cinema Studio for filmmakers directing AI films, and Marketing Studio for marketing and advertising teams. Mashrabov said the money goes to hiring, product and compute, noting that one minute of video is like processing 60,000 words, and that he expects video AI to become “much more deeply embedded in everyday marketing and creative workflows.”
For agencies and brand teams, that last point is the practical one. If even part of the 390 Fortune 500 relationships Higgsfield cites means in-house teams using a self-serve video tool, then some share of campaign content, social variants and internal work is already being made outside the agency and production relationship. That changes what a production partner is hired for. Volume work moves in-house or to a platform; what remains is direction, craft, consistency across a campaign, and responsibility for clearance. The Meta glow-up row in October showed what happens when a platform’s defaults touch a brand’s ads, and the same question of who answers for the output applies to any tool sitting between a model and a client.
There is also a dependency question. An interface layer is only as stable as the models behind it, and those models are licensed from providers whose terms, pricing and training-data exposure can change. Runway’s model router in July points the same way: the interface is becoming the product, and the model underneath is becoming interchangeable. For a production, that means the vendor contract, not the model card, is where provenance and indemnity actually get settled.
Limits worth stating: annualised revenue is a run-rate the company reports, not audited annual income, and the user count does not say how many are paying or retained. A quadrupled valuation in eight months also says as much about investor appetite as about durability. Watch whether Higgsfield starts publishing clearer terms on training data and indemnity for enterprise customers. With 390 of the Fortune 500 named as customers, legal departments will start asking for them.