[SPONSORED] Generating an Interactive World Will Be as Intuitive as Sharing a Video on TikTok, Says Tripo AI Co-Founder

Interview with Dr. Yanpei Cao, Chief Scientist, Tripo AI

As the big names in AI have come to dominate the headlines and public discourse in 2026, one company has been quietly carving out a niche for itself in the field of AI-generated 3D assets for spatial computing and interactive entertainment. In June Tripo AI announced that the company had not only secured $200M in its latest funding round, but that they were actively developing a world model architecture dubbed Project Eden.

“Tripo was never intended to be just an efficiency tool for game studio artists,” explained Dr. Yanpei Cao, chief scientist at Tripo AI. “When we founded Tripo in 2023, the vision was always to build the underlying engine for the next generation of interactive user-generated content (UGC) platforms, infrastructures that would let creators generate and share an interactive 3D world as seamlessly as people upload a video to TikTok today.”

“However, from a first-principles perspective, you cannot compute the dynamics of a world if you haven't mathematically defined the objects inside it,” Dr. Cao noted. “In 2023, high-quality 3D asset generation was the necessary foundational step. We dedicated our resources to solving that spatial equation first, which allowed us to immediately revolutionize existing production pipelines while building the bedrock for our future systems.”

Tripo’s AI suite lets users generate production-ready 3D assets from text or images within seconds. The tools are based on a proprietary, native 3D architecture built entirely in-house, rather than relying on 2D video models or external APIs. “Today, that 3D technology has matured and is actively transforming studio workflows,” said Dr. Cao. “But if you position AI strictly as an efficiency tool for 3D modeling, the commercial ceiling is capped. Our overall goal is to serve as the foundational infrastructure that transforms the entire interactive creation paradigm.”

Enter Project Eden. “LLMs predict the next word, and pure video models hallucinate the next 2D frame, but a world model must understand the physical environment and adapt as it changes in response to user actions, agent behavior, and time.”

By now, the research community's focus had finally caught up with the founders’ original vision. “In our view, world models aren't a separate track, but a fundamental piece of the broader technological puzzle we have been solving since day one,” Dr. Cao maintained.

“Ultimately, we aren't just trying to build another asset generator. We are building the foundational infrastructure for interactive 3D content. Using natural language and high-level semantic intent, users will be able to orchestrate highly interactive worlds.”

Support nonprofit student journalism. Donate to the ‘Prince.’ Donate now »

While Project Eden remains in deep, active research, its initial milestones are already paving the way for new consumer applications. This decoupled architecture is the engine behind an upcoming platform Tripo conceptualizes as an “interactive TikTok,” letting users play bite-sized interactive experiences instead of just passively scrolling through flat videos.

“Our research on world models remains ongoing and our core vision is to empower every individual to architect interactive, complex worlds with the ease of natural expression,” Dr. Cao told the ‘Prince.’ “All of us at Tripo are very excited to see how users leverage tools like Project Eden to bring their creative ideas to life.”