Fal raises $1.65 billion in two funding rounds, valuation doubles in half a year: too many customers, not enough computing power.

Beating AI News Flash: AI inference platform fal has announced the completion of Series E and F funding rounds, totaling $1.65 billion. Among them, the April Series E round was led by Sequoia Capital, raising $350 million at an $8 billion valuation; the October Series F round was led by Bessemer, raising $1.3 billion at a doubled valuation of $16 billion, with Nvidia continuing to participate.

fal primarily provides inference services for image, video, and audio generation models. Developers do not need to rent GPUs or deploy models themselves; they can call models such as Google Nano Banana and MiniMax through fal's API. Companies like Adobe, Canva, and Shopify are also its clients. In recent months, fal has also begun collaborating with model manufacturers to improve generation quality and speed through post-training and inference optimization.

This business is growing rapidly. fal's annualized revenue rose from approximately $200 million in October 2025 to approximately $800 million in September this year.

But doing too well also brings trouble. Founder Burkay Gur revealed that customer demand once exceeded computing power supply this summer, and the company even had to throttle customers. This funding will mainly be used to procure more computing power and expand the team.