Humans are visual creatures. Images and video are the most immersive forms of content. It’s no accident that more than 80% of internet traffic is video, that social platforms are becoming image- and video-first, and that video games and movies are the largest categories of consumer spend.

Generative media will be even bigger. At Sequoia’s inaugural AI Ascent event in 2023, Jensen Huang made the provocative prediction that “Every pixel will be generated, not rendered.” Today, that dream appears closer than ever: frontier video and image models have crossed the uncanny valley, and the first compelling use cases of generative media are starting to emerge across advertising, cinema, storytelling and more.

Projects that once demanded years of work and $100M budgets can now be explored much more quickly and affordably, opening the door to new creative possibilities. We’re seeing generative media begin to transform familiar media use cases: digital ads, viral TikToks, short films, and micro dramas. There will also be wholly new experiences created that we can’t even begin to imagine, from education to personalized media to generated games. The doors to building with creative AI are wide open to anybody with a computer. 

fal has built the leading platform for enterprises and developers to build with generative media models. Video models are compute-intensive and finicky to work with, and creating wonderful outputs requires excellence on multiple levels. fal offers AI creatives exactly what they want for this exciting but strange new paradigm, including 400+ models available on-demand across open- and closed-weight (including models like OpenAI Sora and DeepMind Veo), Day 0 support for new model releases, incredibly fast inference speeds, an ergonomic developer API, an advanced playground UI, and enterprise features around model fine-tunes, styles and collaboration.

fal’s customers are the tastemakers of generative media, with use cases ranging from e-commerce (Shopify) to creative suites (Adobe and Canva) to AI-native platforms (Perplexity) to millions of individual developers. While the momentum behind the business has been staggering, we are even more excited by the quality and caliber of teams currently experimenting on fal, creating immersive new education experiences, virtual pets, indie animation studios, and more. If even a small subset of these explorations make it to production, the world will be a dramatically more colorful, entertaining place.

We are delighted to partner with co-founders Burkay Gur (Coinbase ML) and Gorkem Yurtseven (Amazon) and Head of Engineering Batuhan Taskaya (the youngest-ever Python core developer and maintainer). The team is spiky up and down the platform stack, from having one of the best kernel and compiler inference teams in the world, to nurturing model provider relationships with finesse, to grassroots devrel. Their early bet on generative media shows up across their relationships with model providers, infrastructure performance and ability to dream with creators.

We are at the beginning of a compute explosion in generative video. As the generative media wave accelerates, fal is the inference platform powering the future of AI-first creativity. The team is growing fast to keep up with that demand, and we at Sequoia are proud to lead their Series D.