fal Announces $1.65B Across Series E and F for Generative Media Inference, Led by Sequoia and Bessemer with NVIDIA

Background and pain point

Image, video and audio generation models are moving into production pipelines in film, advertising, e-commerce and games, but their inference is expensive and latency-sensitive, and many teams do not want to run their own GPU clusters. fal is an inference platform for generative media that hosts third-party and in-house models behind an API. On November 3, fal announced on its official X account and blog a combined $1.65B across Series E and Series F.

What is new

According to the fal blog, the Series E was led by Sequoia Capital and closed earlier this year, and is announced together with the Series F, led by Bessemer Venture Partners with continued investment from NVIDIA. fal says the funding covers two things: the inference capacity to run models at scale and a bigger team, plus new research in 2027. fal also post-trains partner models to make them faster and better.

Numbers

This is a funding story, so there are no benchmarks. fal did not disclose valuation or per-round amounts. Bloomberg reported (as relayed by TechCrunch and SiliconANGLE) that the Series F was $1.3B at a $16B valuation, double the $8B valuation of the April Series E, and that annualized revenue passed $800M earlier this year. Those figures are media reports, not confirmed by fal.

What it means for developers

The most direct effect is capacity: per Bloomberg, CEO Burkay Gur said demand was so high over the summer that fal rate limited customers, and the new money goes first to compute. If you call Seedance, Kling, Veo or FLUX through fal, watch for changes in rate limits and pricing, and compare its cost against self-hosted inference or other hosted platforms.