- Founded
- 2021
- Headquarters
- San Francisco, CA
- Latest Round
- Series D
- Est. Valuation
- ~$4,500M
Investment Thesis
Fal is the fastest generative AI inference platform for developers, hosting image, video, and audio AI models at scale. Founded in 2021 by Burkay Gur (former Coinbase ML leader) and Gorkem Yurtseven (ex-Amazon developer), the company provides the infrastructure layer for multimodal AI that powers applications from leading companies.
Fal has surpassed $200 million in revenue as of late 2025, serving customers including Adobe, Shopify, Canva, and Quora. The platform specializes in real-time generative media, enabling developers to build personalized content experiences at scale. With three fundraises in 2025 alone and a $4.5B valuation, Fal has emerged as the leading infrastructure provider for the generative media era.
Running AI inference at scale is slow, expensive, and operationally painful — GPU availability is unpredictable, cold start times kill user experience, and optimizing models for production requires deep infrastructure expertise most teams lack.
Fal provides serverless AI inference infrastructure that handles model serving, GPU orchestration, and optimization automatically — giving developers sub-second cold starts and production-grade performance without managing any infrastructure.
Powering AI inference for thousands of developers and companies building image generation, video AI, and LLM applications; known for best-in-class latency on popular open-source models like SDXL and Whisper.
AI inference demand is growing faster than training demand as models move from research to production, and the complexity of GPU infrastructure management is the biggest bottleneck preventing AI applications from scaling.