AI Video Startup Higgsfield Raises 400 Million at a 5.4 Billion Valuation
DST, Goldman Sachs, Liberty Global and Intel backed the round, more than quadrupling the company's January mark. Revenue reportedly went from 200 million to 500 million annualised in six months.
Higgsfield has closed a 400 million dollar round at a 5.4 billion dollar valuation, according to Financial Times reporting summarised by Techmeme. The investor list runs to DST, Goldman Sachs, Liberty Global and Intel among others. In January the company was valued at 1.3 billion, so this is more than a quadrupling in seven months. Reports in June had it in talks at a 5 billion pre-money valuation while hitting a 500 million dollar annualised revenue run rate, up from around 200 million at the end of 2025.
For anyone who has not come across it: Higgsfield makes tools for generating and editing video with AI, aimed less at film studios and more at the enormous middle market of people who need a steady supply of short video, meaning social media teams, small agencies, e-commerce sellers and marketers. Annualised run rate, worth translating, means the most recent month’s revenue multiplied by twelve. It is a forward-looking estimate rather than money already in the bank, and it flatters fast-growing companies.
The number that actually matters in that paragraph is not the valuation, it is the revenue. AI video has spent two years producing spectacular demos and very little paid usage, mostly because the output was unpredictable and expensive per second. A jump from 200 to 500 million in half a year suggests something shifted from novelty to routine purchase. The likeliest explanation is unglamorous: not feature films, but product clips, ad variants and social posts, where the bar is “good enough and forty times cheaper” rather than “indistinguishable from a film crew”.
Keep expectations calibrated, though. Valuations in this corner of the market are being set by competition among investors as much as by business fundamentals, and a fivefold markup in seven months is a bet on continued growth, not a reward for past growth. The category is also brutally crowded, and the underlying models keep getting cheaper, which is good for customers and hard on anyone whose product is mostly a wrapper around them.
What this means for you: if you make video for work, the tools in this category are now good enough to be worth an afternoon of your time, particularly for the repetitive end of the job like resizing, variants, subtitles and B-roll. Try one before you commit, and test it on your actual footage rather than the demo prompts. If you do not make video, the thing to take away is a market signal: money is flowing into applications now, not just into model labs, which usually means the tools you can actually buy are about to get noticeably better. And if you watch a lot of short-form video, quietly assume that a growing share of what scrolls past you was assembled rather than filmed.
Sources
Source: https://www.ft.com/content/719c8108-f4ae-466b-80f4-96f26558d642
"Kidney Disappointment" and the Papers That Slipped Past Review
A Google Scholar search for a phrase that should not exist returns about 189 results. It is a small, funny window onto how AI-paraphrased text is getting into peer-reviewed journals.
