Over the summer, demand for fal’s image and video models outran the hardware behind them, and the company had to rate-limit customers. It now has $1.3 billion more to spend on that problem. The San Francisco startup closed a Series F led by Bessemer Venture Partners that Bloomberg reported values it at $16 billion, twice its valuation in April.
Nvidia, Sound Ventures, HOF Capital and T. Rowe Price also took part, SiliconANGLE reported. Fal disclosed the round on Tuesday alongside a Series E led by Sequoia Capital that closed earlier this year. Fal said the two rounds total $1.65 billion combined, according to SiliconANGLE and TechCrunch. The company said the money funds work across its platform in 2027 and called out continued investment from Nvidia.
“We hit a point in summer where we were not able to serve our customers and had to rate-limit them,” co-founder and CEO Burkay Gur told Bloomberg, in a quote carried by SiliconANGLE. Gur said the raise is as large as it is because serving that demand takes more compute. Fal also plans to hire beyond its current staff of 185.
Media models need a different kind of serving
Fal runs image, video and audio generation models that developers reach through a single API. In a post explaining the investment, Bessemer’s partners argued that the serving stacks built for chatbots don’t carry over: “LLMs produce one token at a time. Image and video models refine a whole frame over many passes, so they need memory, compute, and latency handled differently.” They wrote that Adobe, Canva, Shopify, World Labs and Amazon have moved production volume to fal.
The company has also started improving the models it hosts. Bessemer said fal Research released H3 Max, a post-trained version of MiniMax H3, in August, and partnered with Black Forest Labs on Flux 3 Max last month. Some of those model arrangements pay fal licensing fees or a share of revenue, according to Bloomberg. Gur said pretraining a model of fal’s own isn’t on the near-term agenda, though he didn’t rule it out.
Big rounds, thin margins
Fal passed $800 million in annualized revenue earlier this year, a person familiar with the matter told Bloomberg. Its new valuation sits just behind rival inference provider Fireworks AI, which SiliconANGLE said raised $1.5 billion at a $17.5 billion valuation in July. TechCrunch counted Baseten and Modal among the other fast-growing inference companies raising at higher prices, and said several people familiar with their financials described gross margins as thin because compute is expensive to buy or lease.
Fal isn’t the only AI company paying up for capacity. Anthropic said in an amended IPO filing on Monday that it spent $5.0 billion on short-term compute last quarter, which costs more than capacity bought under long-term contracts. On the supply side, AMD reported Tuesday that its data center sales nearly doubled from a year earlier.
Bessemer partner Talia Goldberg, who the firm said has known Gur and co-founder Gorkem Yurtseven for more than a decade, is joining the board of Features & Labels Inc., fal’s formal name. With this round, investors have put more than $1.9 billion into the company, by SiliconANGLE’s count.


WebProNews is an iEntry Publication