Moonlake AI

About Moonlake AI

Moonlake AI is a San Francisco-based applied research lab developing multi-modal reasoning models that allow users to create interactive content, such as games and simulations, using natural language. The company emerged from stealth on October 1, 2025, with a $28 million seed funding round led by AIX Ventures, Threshold, and NVIDIA Ventures, with participation from notable angel investors including the founder of YouTube, Steve Chen, and the founder of AngelList, Naval Ravikant.

The company was founded by Fan-Yun Sun and Sharon Lee, both researchers from the Stanford AI Lab. Sun, who earned his PhD in Computer Science at Stanford, previously worked as a researcher at NVIDIA, where he led efforts in generating large-scale 3D worlds for training AI agents. Lee was also pursuing a PhD at Stanford, focusing on generative AI, computer graphics, and creating models that can reason about space and interactivity. Their goal is to reduce the complexity and cost of creating interactive worlds, a process that has traditionally been limited to large studios or developers with specialized skills.

Moonlake AI's platform enables creators, developers, and designers to build and modify interactive 2D and 3D environments in minutes without needing coding or 3D modeling skills. The technology integrates multi-modal reasoning for spatial planning, program synthesis for game logic, and a real-time diffusion model for visual styling. This allows users to conversationally edit physics, add rules, and introduce AI agents into the environment. The platform is aimed at a range of clients, including game developers, design studios, IP holders, and researchers. Initial applications include rapid prototyping for games, user-generated interactive worlds, and creating simulation environments for training robotics and embodied AI agents.