Reactor raises $59M to power real-time generative AI video for developers
What's the deal? Reactor, a San Francisco-based startup building a platform for real-time generative AI video, has raised $59M in a Series A round. Lightspeed Venture Partners led the investment, with WndrCo, Amplify Partners, Sky9 CapitalDealroom has a profile for this one. Try Dealroom →, and FPV VenturesDealroom has a profile for this one. Try Dealroom → also participating.
Founded in 2025 by Alberto Taiuti and Bryce Schmidtchen, Reactor provides a unified software development kit and API that lets developers build real-time interactive AI applications at scale. It positions itself as a critical software layer between AI model labs and the developers who want to use those models.
The company plans to use the funds to expand operations internationally and within the US, grow its GPU capacity for real-time processing, and invest in marketing.
Why now? Generative AI video has leapt from research demos to production-ready tools in a remarkably short time. As major model labs release increasingly capable video generation systems, a gap has opened for infrastructure that helps developers actually integrate these models into live, interactive applications — exactly where Reactor sits.
The timing also reflects growing enterprise demand for dynamic, user-interactive media generated on the fly, rather than pre-rendered content.
What could go wrong? Real-time AI video processing is enormously GPU-intensive. Scaling that infrastructure while keeping costs manageable is a serious challenge, especially as demand spikes. Reactor also faces the risk that the large model labs it sits between — companies like OpenAI, Google, and others — could build their own developer tooling, squeezing out the middleware layer.
Competition in the generative AI infrastructure space is fierce, and the technology is evolving so quickly that today's differentiator can become tomorrow's commodity.
The signal: Lightspeed leading a $59M Series A for a company barely a year old speaks to how aggressively top-tier VCs are betting that the real bottleneck in generative AI video isn't model capability but developer-facing infrastructure. With WndrCo — the firm co-founded by Jeffrey Katzenberg — also on the cap table, the round signals conviction that real-time AI video is crossing from tech demo to entertainment and media production at scale.
Read more: thesaasnews.com
