Disclosure: Higgsfield is a SaaStr Fund portfolio company. We invested early, and we’ve been customers since the early days.
Higgsfield crossed $1 billion in annualized revenue run rate in September. Alex Mashrabov just sent me the October number: $1.35 billion. That’s $350 million of new run rate in about a month. Most companies at that scale would be happy to add $350 million in a year.
The company is now targeting $2 billion by the end of 2026. UBS also put Higgsfield on its AI IPO watchlist, alongside Replit, Perplexity, Databricks, Anduril and Canva. Alex reads that list as putting Higgsfield at #3 by reported run rate among independent, privately held AI-native companies, and as the largest of the generative AI-native names on it.
We’ve covered a lot of fast AI ramps on SaaStr this year: Cursor, Lovable, Replit, ElevenLabs. Higgsfield is the one where the dollars added each month are still going up past $1 billion.
The short version:
- $1B to $1.35B in one month. That’s 35% month-over-month growth at $1B+ in run rate.
- The monthly dollar add went up at every step. About $60M a month in the first half of 2026, $100M a month in early summer, $300M in September, $350M in October.
- $2B by year-end needs no further acceleration. Holding October’s dollar pace for two more months gets there.
- Agencies and enterprise customers drive most of the revenue. Individual creators are a minority.
- The team stayed small. About 150 people at $500M, and cash-flow positive since May.
1. The Run-Rate Ladder: $20M to $1.35B in About 14 Months
These are the company’s own reported figures:
- August 2025: roughly $20 million in annualized revenue
- End of 2025: roughly $200 million
- February 2026: Forbes reports Higgsfield has crossed $300M in annualized revenue run rate
- May 2026: ~$500 million, when Alex came on stage at SaaStr AI
- August 2026: $700 million, alongside the Series B
- September 2026: $1 billion
- October 2026: $1.35 billion (per Alex)
- Target, December 2026: $2 billion
In February, Mashrabov said on the sidelines of Web Summit in Qatar that Higgsfield was aiming to reach a $1 billion annual run rate by the end of the year. It hit that in September, reaching its year-end target three months early. The $2B goal comes from a team that already beat its own 2026 forecast by a full quarter.
2. At SaaStr AI in May: ~$500M, ~150 People, No Foundation Model
Alex was on stage with me at SaaStr AI in San Mateo in May, with Higgsfield at about $500M in run rate. The agenda billed the session at $300M because it went to print months earlier. October’s $1.35B is 2.7x that number, five months later.
What Alex laid out in May:
- They chose not to train their own foundation model. Every AI company is racing to train their own foundation model. Alex Mashrabov did the opposite. Higgsfield spent on the workflow, model routing and its own inference stack.
- Agencies were already most of the revenue. About 70% at the $300M stage.
- About 150 people, roughly 60 in core engineering, and cash-flow positive.
- The agentic product shipped that same month. Following the company’s Supercomputer rollout in May 2026, users of Higgsfield’s agentic products have grown 42-fold in just three months, driving more than 20 million content generations per month. The monthly dollar adds jumped from ~$100M to $300M+ over the following quarter.
The company you see at $1.35B is the same one that was on stage in May, with the same model strategy, the same buyer and roughly the same team shape.
3. The Monthly Dollar Add Went From ~$60M to $350M
| Period | Run rate | Added per month |
|---|---|---|
| Jan to Jun 2026 | $200M → $500M | ~$60M |
| Jun to Aug 2026 | $500M → $700M | ~$100M |
| Aug to Sep 2026 | $700M → $1B | ~$300M |
| Sep to Oct 2026 | $1B → $1.35B | ~$350M |
October’s add alone is bigger than Higgsfield’s entire run rate at the end of 2025.
At $1B, most companies see percentage growth slow while dollar adds hold flat, and that counts as a good outcome. At Higgsfield, both numbers rose from August to October.
4. Agencies and Brands Are the Buyer
Higgsfield started out looking like a creator tool. The revenue says otherwise. Alex told us on SaaStr that roughly 70% of Higgsfield’s revenue around the $300 million run-rate stage came from agencies. The average Higgsfield customer was spending around $1,000 per year, approximately five times the roughly $200 figure it cited for Canva.
The enterprise side is growing faster still. Revenue under contract from business customers has grown tenfold since June, driven in part by demand from direct-to-consumer businesses producing AI-generated advertising. At the Series B, the company reported above 30 million users and work produced for 390 Fortune 500 companies.
The product was rebuilt around that buyer. In December 2025, Stripe flagged Higgsfield as a top-10 fastest-growing account on its platform as it crossed a $200M run rate and launched Team and Enterprise plans.
That matches the bar I use for agentic B2B investments: deal sizes have to pull from budget beyond the software line. Higgsfield is spending down production budgets. Mashrabov also claims the platform cuts production costs for businesses by up to 97 percent. The money is coming out of shoots, editors and agency retainers.
5. Gross Margin Turned Positive in Early 2026, Cash Flow in May
Video generation started with bad unit economics. Positive gross margins only since early 2026 mean that last year, Higgsfield paid more for compute and model providers per generated video than it took in. Then Higgsfield turned cash-flow positive in May 2026.
The infrastructure underneath is large. Higgsfield said it processes 12 trillion tokens across 150 million generations every month. It uses NVIDIA Dynamo to deploy and scale custom inference workloads on its own GPU infrastructure.
The sequence: demand first, then positive gross margin, then positive cash flow, then the $400M raise.
6. The Capital Followed the Numbers
- January 2026: Series A extension at $1.3B, on $200M of run rate
- August 2026: USD 400 million at a USD 5.4 billion valuation, led by DST Global. At the time of the round, Higgsfield reported $700M in annualized revenue, implying a roughly 7.7x annualized revenue multiple.
- October 2026: Nvidia and TCV extended the Series B. The company did not disclose the amounts or its new valuation. It is Nvidia’s first investment in Higgsfield.
Against $1.35B of run rate, the August price is about 4x.
7. Run Rate Here Means the Last 4 Weeks Times 13
Higgsfield discloses its method. Mashrabov said Higgsfield takes revenue from its latest four weeks and multiplies it by 13. That’s current pace, not a trailing year of collected revenue.
Every fast AI company reports some version of this, and you should read it for what it measures. According to Mashrabov, annual subscriptions account for about 40 percent of revenue. The other 60% is credits, top-ups and usage beyond plan limits, which can move month to month.
What gives the number weight for me: eight reported checkpoints with no step back, positive cash flow since May, and a revenue mix moving toward contracted enterprise dollars.
8. Five Decisions Behind the Curve
- Higgsfield followed the buyer who was paying. It expected creators and found agencies and brands, then shipped Team and Enterprise plans in December 2025.
- It priced against production budgets. About $1,000 per customer per year, roughly 5x Canva’s figure, because the alternative is a video shoot.
- It skipped training a foundation model. It routed across models and put the money into workflow and inference.
- It fixed gross margin before raising big. Negative unit economics in 2025, positive gross margin in early 2026, cash-flow positive in May, then $400M in August.
- It published its run-rate math. Four weeks times 13, stated on the record.
The Math on $2B by December
From $1.35B, Higgsfield needs about $650M more in run rate by year-end. October added $350M. Two more months at $300M to $350M each gets there.
If they hit it, Higgsfield will have gone from about $200M to $2B in run rate in calendar 2026, cash-flow positive since May, with about 150 people at the halfway mark.
