Fal’s valuation doubles to $16bn, Bloomberg reports

T. Rowe Price, Sound Ventures and Nvidia also took part. Fal revealed the round together with a $350 million Series E, led by Sequoia and closed in April, which brings the two rounds to $1.65 billion. Tech.eu adds that the funding goes to Fal’s inference servers, its team and its generative media technology. Bloomberg’s report puts Fal’s yearly revenue run rate, meaning current sales scaled up to a year, above $800 million since earlier in 2026.

Inference is what Fal sells: once a model is trained, a server runs it every time someone asks for a clip or an image, and Tech.eu says advertising, e-commerce and marketing teams are among its customers. TechCrunch names Baseten, Fireworks AI and Modal as other inference firms growing rapidly. Several people familiar with such firms’ finances told TechCrunch their gross margins stay thin, since computing power is expensive to buy or lease.

We hold no Fal price, but the list rates show what this market charges at 720p. Veo 3.1 Lite was $0.05 a second and Fast $0.10 at our 4 November check, Wan 3.0 was $0.083 on 30 October and Kling 3.0 with audio $0.126 on 31 October (our price sheet).

On those rates the model sets the bill: 20 eight-second clips cost about $8 on Veo 3.1 Lite and about $20 on Kling 3.0 with audio.