Moonlake AI launches with $28M to let anyone build interactive game worlds in minutes

Building games and interactive worlds has long been a privilege of large studios or highly skilled developers. For most creators, the process is either too slow, too expensive, or simply out of reach. That gap is what Moonlake AI is aiming to close.

The San Francisco–based AI startup emerged from stealth today with $28 million in seed funding led by AIX Ventures, Threshold, and Nvidia Ventures, with participation from more than a dozen well-known AI researchers and founders. Backers include YouTube co-founder Steve Chen, AngelList’s Naval Ravikant, GANs inventor Ian Goodfellow, and executives from Hugging Face, DeepMind, Stability AI, and OpenAI.

Moonlake AI is developing reasoning models that let anyone describe a game or simulation in natural language and see it transformed into a working interactive scene within minutes. The platform compiles sketches, ideas, or even moods into playable 2D and 3D environments that can be tweaked instantly. “Game design is about to become 10X better with Moonlake’s AI platform,” said Shaun Johnson. “This will enable consumers and professional game designers to create high-grade games using natural language.”

Moonlake AI Emerges From Stealth With $28M Backing From YouTube Co-Founder, Others

Steve Chen sees it as a parallel to the platform he built two decades ago. “I’m personally excited by the mass potential that Moonlake AI is working to unlock. Similar to YouTube opening up a platform for the world to watch and share videos, I look forward to Moonlake AI doing the same for interactive content,” Chen said.

Behind the scenes, Moonlake’s tech blends multimodal reasoning with program synthesis and simulation layers. It handles spatial layout, physics, and agent behaviors, while a diffusion model reskins the world in real time. That means creators can sketch a mechanic or describe a scene and instantly have something playable—whether it’s for a game prototype, robotics simulation, or classroom experience.

The company was founded by Fan-Yun Sun and Sharon Lee, both researchers out of Stanford’s AI Lab. Sun previously worked at Nvidia on generating large-scale 3D environments for AI training, while Lee specialized in combining generative models with 3D engines to create richer interactivity. Their shared vision is to lower the barrier to world-building for both professional developers and everyday creators. “Too many great worlds never made it past the sketchpad,” Lee said. “It is not for lack of imagination, but for lack of the right tools.”

Moonlake’s approach arrives at a time when budgets for blockbuster games often stretch into the billions and prototyping alone can take years. By cutting down the cost and time it takes to build interactive environments, the startup is betting that it can broaden who gets to create—not just large studios, but indie developers, educators, researchers, and the billion-plus individual creators who want to experiment with interactive content.

The platform is currently in private preview, with early demos showing instant world-building from text prompts and live editing inside simulations. For now, interested developers and creators can request early access through the company’s site.