A San Francisco-based startup called Fal, which provides hosting and operation of AI models for video, image, and audio generation, has secured $1.3 billion in a funding round spearheaded by Bessemer Venture Partners. Nvidia, Sound Ventures, and T. Rowe Price also participated in the round. According to Bloomberg, the deal places Fal's valuation at $16 billion—twice the $8 billion figure from April. The company simultaneously disclosed both its Series E and Series F rounds, revealing that a total of $1.65 billion was raised across the two. Sequoia had led the earlier round.

This newest round—the Series F—arrives as demand surges for inference services, which involve running an already-trained AI model to produce outputs. Customers relying on open-source models are a particularly strong source of this demand. Bloomberg's reporting noted that Fal exceeded $800 million in annualized revenue earlier this year. Other companies in the inference space, such as Baseten, Fireworks AI, and Modal, have likewise been expanding quickly and securing large funding rounds at elevated valuations. However, according to multiple individuals familiar with the financials of inference companies, gross margins across the sector remain slim due to the steep expense of purchasing or leasing compute.

Fal was established in 2021 by Burkay Gur—previously a machine learning leader at Coinbase and an engineer at Oracle—and Gorkem Yurtseven, who had worked as a developer at Amazon.