Fal has raised $1.3 billion in a new funding round, doubling its valuation to $16 billion from $8 billion in April. The San Francisco-based startup hosts and runs AI video, image, and audio models, and its latest financing arrives as demand for AI inference services rises.
Two rounds bring in $1.65 billion
The company announced its Series E and Series F rounds at the same time, reporting that it raised a combined $1.65 billion across both. Sequoia led the earlier round, while the latest, Series F, was led by Bessemer Venture Partners.
Nvidia, Sound Ventures, and T. Rowe Price also participated in the latest round. Bloomberg reported the new $16 billion valuation, up from the $8 billion figure in April.
The combined total and the latest round describe different pieces of the fundraising: $1.65 billion covers both rounds, while $1.3 billion was raised in Series F. Taken together, the announcements show how much capital Fal has attracted across the two financings.
Inference demand is driving investor interest
Inference is the step in which an already-trained AI model generates outputs. Fal provides the services needed to host and run models, including video, image, and audio systems. The article says demand for inference services is soaring, particularly from customers relying on open-source models.
That demand has also coincided with rapid growth and large funding rounds at other inference providers, including Baseten, Fireworks AI, and Modal. The pattern points to growing interest in the companies that make trained AI models usable by customers.
Revenue is another sign of Fal’s momentum. Bloomberg’s report noted that the startup surpassed $800 million in annualized revenue earlier this year. That figure is annualized revenue, rather than a description of revenue earned over a full year.
High compute costs complicate growth
Rising demand and large valuations do not tell the whole financial story. People familiar with inference companies’ financials told TechCrunch that gross margins remain thin across the sector because compute is expensive to buy or lease.
For providers such as Fal, the costs of the computing capacity needed to run models can weigh on the amount left after delivering services. The source does not give Fal’s own gross margin, so the reported constraint applies to inference companies generally, rather than establishing a specific margin for Fal.
The sector therefore faces a practical challenge alongside its growth: serving more customers can support revenue, while the compute required to do so remains costly. Funding can support expansion, but the article’s account of thin gross margins underscores that demand alone does not settle the economics.
A startup built around AI models
Fal was co-founded in 2021 by Burkay Gur and Gorkem Yurtseven. Gur was a former Coinbase machine learning leader and Oracle engineer; Yurtseven was previously a developer at Amazon.
The founders’ company now operates in a market where customers are seeking ways to run AI models, including open-source ones. Its latest funding and valuation reflect investor backing in that environment, while the cost of compute remains a central factor for inference services.