Fal

Training AI models grabs the headlines, but running them pays the bills. Fal, a San Francisco startup that hosts and runs AI video, image and audio models, just proved how hot that business has become by doubling its valuation in under seven months. Quick Facts What Fal Actually Does Fal provides infrastructure for inference, the process of running an AI model that’s already been trained so it can generate outputs. Developers and companies use Fal to run AI video, image and audio models without managing the underlying hardware themselves. Demand is especially strong from customers relying on open-source models. The Growth Behind the Valuation Fal announced its Series E and F rounds together, with Sequoia leading the earlier one. Bloomberg reported that the startup passed $800 million in annualized revenue earlier this year. That kind of growth in a five-year-old company explains why investors keep raising their bets. A Crowded, Fast-Growing Category Fal isn’t alone. Inference providers including Baseten, Fireworks AI and Modal have also grown quickly and raised large rounds at higher valuations. As more companies move AI features from experiments to production, the cost of running models every day, not just training them once, becomes a major spending line. The Catch: Thin Margins Strong revenue doesn’t automatically mean strong profits. According to several people familiar with inference companies’ financials, gross margins across the sector remain thin because buying or leasing compute is so expensive. In other words, these startups collect plenty of revenue but pay much of it to chip and cloud providers. Meet the Founders Burkay Gur, a former machine learning leader at Coinbase and an ex-Oracle engineer, co-founded Fal with Gorkem Yurtseven, who previously worked as a developer at Amazon. Why Nvidia’s Involvement Matters Nvidia’s participation reflects how closely chipmakers tie themselves to the companies that rent out their hardware. Inference startups are major buyers of GPUs, so a backer like Nvidia benefits when they grow. What It Means for Developers and Businesses What Readers Are Asking How much did Fal raise? $1.3 billion in its Series F. What is its valuation? $16 billion, double April’s $8 billion. Who led the round? Bessemer Venture Partners. What is inference? Running a trained AI model to produce outputs. How much revenue does Fal have? Bloomberg reported more than $800 million in annualized revenue earlier this year. What to Watch Next Look for further inference funding rounds, signs of improving margins and whether Fal expands beyond media models into other AI workloads. Source: TechCrunch