SAN FRANCISCO — Goldman Sachs Alternatives does not put its name on novelties. When the firm joined DST Global in leading Higgsfield’s $400 million Series B, announced Sunday, the signal was not that AI-generated video is interesting. It was that AI-generated video belongs in the same sentence as enterprise infrastructure. That is a different claim than anything the AI video industry has made for itself, and the $5.4 billion valuation attached to the round is the market’s answer to whether it believes it.
The company disclosed the round alongside metrics that supply the basic case: $700 million in annualized revenue, 30 million users across 200 countries, and 390 Fortune 500 companies now using its tools for marketing and creative production. Alex Mashrabov, who left Snap to found Higgsfield in 2023, said the capital would go toward hiring and compute capacity, in roughly equal proportion. Video is one of the most compute-intensive domains in AI, Mashrabov noted in the company’s announcement. A minute of generated video requires processing roughly equivalent to 60,000 words of text. At $700 million in revenue, how much survives after compute costs is not something Higgsfield has disclosed.
The valuation tells a sharper story. Higgsfield was worth $1.3 billion eight months ago. The $5.4 billion figure represents a quadrupling over a period in which every major investor in artificial intelligence has been competing to write checks into any company with credible enterprise metrics. DST Global and Goldman Sachs Alternatives are sophisticated enough to know that context. Whether the $5.4 billion reflects Higgsfield specifically, or the broader AI funding moment it is operating inside, is not a question the funding announcement addresses.
What the round does answer, partly, is how the company built its way to Goldman’s attention. Higgsfield’s Cinema Studio has positioned the product at the high end of the creative market, with AI-generated films that have premiered at Cannes and in New York. No distributor has signed a Cannes-premiered AI film to a theatrical release, but the festival presence functions as a proof-of-concept for the technology’s creative ceiling. It signals that Higgsfield wants its tools associated with directors, not just marketing teams looking for cheap campaign video.
At Cannes, Higgsfield debuted a short film created using its platform, with Mashrabov in the director’s chair alongside the platform’s tools. The film premiered alongside studio-produced features, which is either a marketing exercise or a genuine statement about where AI video quality has arrived. What it accomplished was getting the company’s name into conversations about filmmaking rather than only advertising technology, which matters when enterprise marketing teams are deciding what kind of vendor relationship they want.
The Marketing Studio is where the revenue lives. Synthesia and Runway compete in adjacent territory, with Synthesia targeting enterprise communications and Runway building toward the creative and film market. Higgsfield’s competitive claim rests on compute reliability: that the company has secured enough capacity to serve enterprise video workloads at the scale that 390 Fortune 500 procurement teams require, without the performance degradation that becomes a genuine problem when hundreds of companies are generating campaign material simultaneously.
Valor Capital and Tribe Capital joined DST Global and Goldman Sachs Alternatives in the round. The bet the investors are making is that enterprise video AI is about to pass through the same adoption curve that enterprise document AI entered after ChatGPT appeared in late 2022: slow pilot adoption followed by fast budget deployment once internal teams proved productivity gains. Whether that curve arrives in 12 months or three years is what the check size is implicitly answering.
Mashrabov’s stated priorities, hiring and compute, are honest about where the constraint lies. AI video companies operate under a hardware ceiling that language model companies at comparable revenue do not face in the same proportion. A company generating $700 million in revenue from language model subscriptions spends a materially different fraction on compute than a company generating the same revenue from video. Higgsfield has not disclosed its cost structure, which means the $5.4 billion valuation is built on the ARR trajectory, on sector multiples during an AI investment cycle without precedent in prior markets, and on the assumption that GPU costs will fall fast enough to protect unit economics as the company scales.
The 30 million users represent mostly free-tier customers. The enterprise business, the 390 Fortune 500 clients, is where the $700 million comes from. That implies average enterprise revenue of roughly $1.8 million per client per year, which is not a marginal spend. Marketing budgets at that level do not get renewed without a production argument, and the renewal question is the one Higgsfield’s announcement does not address. ARR trajectory figures, which the company also did not disclose, would answer the retention question more directly than client counts or total valuation.
The round lands in a year of sustained and accelerating AI infrastructure investment. Anthropic’s annualized revenue reached $65 billion in the same period, and Nvidia committed $1.5 billion to SB Energy’s OpenAI data center campus in Ohio, signaling that even chipmakers are becoming infrastructure investors to guarantee demand for their own hardware. Into that environment, Goldman Sachs writing a check for AI-generated video is less surprising than it would have been a year ago. That context is also exactly why the $5.4 billion valuation is difficult to read in isolation from the funding cycle that produced it, as TechCrunch first reported.
Related Posts
