Generative media inference startup fal raises $1.3B at a $16B valuation
Generative media inference startup fal announced it has completed a new $1.3 billion funding round at a $16 billion valuation, with the funds to be used for compute and to expand hiring beyond its existing 185 employees. fal provides a cloud platform that calls image, video and audio generation models through a single API, serving entertainment studios, advertising agencies and game developers; CEO Burkay Gur said that this summer the company had to throttle customers due to surging demand. The round was led by Bessemer Venture Partners, with participation from NVIDIA, Sound Ventures, HOF Capital and T. Rowe Price, bringing the company's cumulative funding to more than $1.9 billion.
Why read this
AI score 62
The link between the size of the funding round and the compute bottleneck can help readers understand the capital and cost structure of the generative media inference sector.
Topics
Body
Generative media inference startup fal today announced it has raised $1.3 billion in new funding at a valuation of $16 billion, according to Bloomberg.
The company will spend the money on computing power and on hiring beyond its current 185 employees.
Fal provides a cloud platform for running image, video and audio generation models, and developers reach all of them through a single application programming interface. Entertainment studios, advertising agencies and game developers are among the customers using those models in production, according to the company.
In-house researchers also post-train some of those models to run faster. Over the past year that work has centered on video, in partnership with the open- and closed-source developers that build the models. Some of the deals pay fal licensing fees or a share of revenue, Bloomberg reported.
Co-founder and Chief Executive Burkay Gur told the outlet the round is as large as it is because demand has surged and serving it takes more compute. “We hit a point in summer where we were not able to serve our customers and had to rate-limit them,” he said. Pretraining a model of fal’s own is not on the near-term agenda, said Gur, who did not rule it out for later.
Annualized revenue at fal passed $800 million earlier this year, a person familiar with the matter told Bloomberg. The valuation puts fal just behind rival inference provider Fireworks AI Inc., which closed $1.5 billion in funding at a $17.5 billion valuation in July. Several people familiar with the finances of inference companies told TechCrunch that gross margins across the sector remain thin because compute is expensive to buy or lease.
Gur started the company in 2021 with Chief Technology Officer Gorkem Yurtseven. His resume includes leading machine learning work at Coinbase Global Inc. and an engineering role at Oracle Corp. Yurtseven was previously a developer at Amazon.com Inc.
Bessemer Venture Partners led the Series F round, and its partner Talia Goldberg is joining the board of Features & Labels Inc., as fal is formally known. Nvidia Corp., Sound Ventures, HOF Capital and T. Rowe Price Group Inc. also participated. Fal described Nvidia as a strategic investor.
Today’s announcement also covered an April round led by Sequoia Capital. That Series E round raised $350 million at an $8 billion valuation, per Bloomberg, and fal said the two rounds come to $1.65 billion combined. Last December, the company raised $140 million at a $4.5 billion valuation.
Investors have now put more than $1.9 billion into the company.
Image: fal