Higgsfield raises $400M led by DST Global at a $5.4B valuation

DST Global led a $400 million Series B into Higgsfield, the AI video and image generation startup, at a $5.4 billion valuation. Growth Equity at Goldman Sachs Alternatives participated as a new investor. It did not co-lead, and it did not lend: this is equity, in a syndicate that also brought in Tribe Capital, Smash Capital, Fifth Wall, Valor Capital, Intel Capital, Liberty Global Tech Ventures, Mirae Asset Capital and NTT DOCOMO Ventures, per the company's announcement. TechCrunch reported the round the same day.

The step up is 4.2x, not 4x, and the base was $1.3B

Higgsfield's Series A priced the company at $1.3 billion, not the $1.4 billion figure that has circulated. From $1.3 billion to $5.4 billion is a 4.2x markup in about eight months. The company was founded in 2023 by Alex Mashrabov, previously an executive at Snap where he ran generative AI, and Yerzat Dulat.

The product surface is two studios plus infrastructure: Cinema Studio for people directing AI films, Marketing Studio for brand and agency teams, and a compute layer the company launched in May 2026 and calls Supercomputer. Named peers in the same market are Synthesia and Runway.

Do the multiple yourself

Here is the number the headlines skipped. At $5.4 billion on a $700 million annualized run rate, Higgsfield is priced at roughly 7.7x revenue, call it 8x.

That is a cheap-sounding multiple for AI, and it is the single most important thing to be careful about, because annualized is not annual. An annualized run rate takes a recent period, usually one month, and multiplies it out by twelve. It is not audited, it is not trailing twelve months, and it does not survive a bad quarter. A company that booked $58 million in its best month can report a $700 million run rate having collected a fraction of that over the past year. The 8x is real arithmetic on a number whose denominator is a snapshot.

Two figures make the snapshot easier to trust or distrust. Higgsfield says it has more than 30 million users across 238 countries and territories, with the US as its largest market, and that it powers visual production for 390 of the Fortune 500. That last one is worth pausing on: 390 out of 500 is 78% of the list. Penetration that broad usually means a lot of small seat-level or pilot spend rather than a few enormous contracts, which is the profile that inflates a monthly snapshot most easily when a promotional month lands.

What the money is for

Proceeds go to R&D, global infrastructure, AI talent and go-to-market, per the release. TechCrunch reports compute capacity is the priority, and Mashrabov has been explicit about why. "Video is one of the most compute-intensive domains in AI," he told TechCrunch. "Just one minute of video is like processing 60,000 words."

On the commercial thesis, Mashrabov's line in the company release is narrower than the usual AI boilerplate: "Every business needs visual content, but creating it at the quality, speed and scale companies demand remains complex and expensive." He expects enterprise adoption of video AI, in his words, "to become much more deeply embedded in everyday marketing and creative workflows."

The takeaway

Two things to do with this. First, recompute the multiple when a real revenue period gets disclosed. If the $700 million annualized figure came off a promotional or seasonally strong month, the true trailing multiple is meaningfully higher than 8x, and the 4.2x markup was underwritten against a number that was never a year of revenue. Second, watch which side of the business grows. Consumer creators are what got Higgsfield to 30 million users, but enterprise is what would justify $5.4 billion, and the 390-of-the-Fortune-500 stat tells you nothing about contract size. Ask for net revenue retention on the enterprise cohort before you treat this price as conservative.