From $1.3B to $5.4B in Seven Months: Higgsfield's $400M Bet on AI Video for Business

Seven months. That is all the time it took Higgsfield to go from a $1.3 billion upstart to a $5.4 billion company — one of the fastest valuation climbs in the history of the generative AI boom. On August 17, the Financial Times reported that the AI video generation startup has closed a $400 million funding round at a $5.4 billion post-money valuation, more than quadrupling the mark it set in January.

The round reads like a who’s-who of institutional capital: Goldman Sachs (through its Equity Growth fund), DST Global, Liberty Global, and Intel backed the deal, with Tribe Capital, Smash Capital, Fifth Wall, Valor Capital, Mirae Asset Capital, and NTT DOCOMO Ventures also participating. For a company that publicly launched its browser-based video platform only in 2025, the speed is staggering. The backstory of how it got here says a lot about where the AI video market is actually heading.

From Snapchat lenses to a $5.4B video engine

Higgsfield was founded in 2023 by Alex Mashrabov, who previously sold his computer-vision company AI Factory to Snap for $166 million in 2020 and then ran generative AI inside the social media giant. He left to start Higgsfield with co-founder and CTO Yerzat Dulat, and their original thesis was contrarian when almost everyone else was racing to train the biggest proprietary video model.

Instead of building one monolithic model, Higgsfield plugged in several external models and concentrated its engineering effort on the workflow around them — letting a user keep reworking, editing, and iterating on a clip rather than generating one take and calling it done. The platform lets users turn text and images into video, and to control camera movements and visual effects without traditional filming equipment. That camera-control focus is what early creative directors told the founders they actually cared about, and it became the product’s signature.

The bet worked. The platform now claims more than 30 million users across 238 countries and territories, with the United States as its largest market.

The real story: enterprises are paying the bills

The consumer footprint is impressive, but the number that unlocked this round is on the revenue line. According to the company, annualized revenue reached roughly $700 million in August — up from around $20 million a year earlier, and up from roughly $200 million at the end of 2025. (Third-party tracker Sacra had estimated $500 million annualized as of June, so the internal figure represents a further acceleration.)

More telling than the headline number is its composition: business customers now account for the majority of revenue, compared with less than a quarter in January. Higgsfield has increasingly pitched itself at marketing teams that need a steady stream of social media content without commissioning every asset through agencies and production companies. Brands on the platform are reportedly producing several videos a day for social channels and advertising, rather than one polished campaign asset per quarter.

Mashrabov told the FT the new capital would help Higgsfield “accelerate our move upmarket” as it targets larger corporate customers. Goldman Sachs, for its part, has previously estimated that the global creator economy could grow from $250 billion in 2023 to $480 billion by 2027 — and the shift toward short-form video keeps pushing brands to produce ever more content across platforms.

A market that is thinning, not crowding

The competitive context makes the round more interesting. Higgsfield competes with OpenAI’s Sora, Google’s Veo, and specialist players like Runway as tech groups race to improve the quality and length of generated clips. But the field has been reshaping fast:

• Runway, Higgsfield’s closest Western rival, raised $315 million at a $5.3 billion valuation in February — but has since pivoted toward world models for robotics and medicine, largely stepping out of the marketing lane.
• In China, Kuaishou’s Kling raised nearly $3 billion at an $18 billion valuation, dominating the domestic market.
• OpenAI and Google bundle video generation into broader model platforms, which makes them competitors and suppliers at the same time.

That leaves the enterprise marketing niche — the segment actually converting views into revenue — surprisingly open for a specialist. Higgsfield’s model-agnostic approach also insulates it somewhat: if a better video model ships next month, the workflow layer can absorb it.

The compute problem behind the numbers

Where is the $400 million going? The company says the capital is earmarked for enterprise products, security, and compute capacity. That last item is the real constraint. Generating video requires dramatically more computing power than producing text or images, making inference cost one of the biggest hurdles facing the entire sector. Running out of compute is a faster way to lose customers than losing a feature race — and video-native growth like Higgsfield’s burns through training and serving capacity at a pace few startups can fund from operations.

The participation of Intel in the round is notable in this light: a chipmaker taking an equity position in one of its potential large customers is a classic way to lock in demand while giving the startup a compute partner with skin in the game.

The open questions

For all the momentum, the round invites scrutiny. A valuation of $5.4 billion against annualized revenue of roughly $700 million implies a multiple of under 8x run-rate — rich but not irrational by 2026 AI standards. The steeper question is durability: can revenue keep compounding at anything close to this rate, or is the $20M-to-$700M climb in twelve months a one-time land-grab as enterprises experiment with AI video? Compute costs could compress margins as volumes scale, and the creative industries are watching the technology’s push from advertising into film production with growing alarm over jobs.

What is not in question is the direction. Video is becoming the default surface for AI-generated content, the buyers are businesses rather than hobbyists, and the money — Goldman, DST, Intel, Liberty Global — has decided the workflow layer is where the value pools. Higgsfield just became the clearest expression of that thesis.