Tripo AI, a developer of interactive 3D foundation models and world models, said today it has closed on a hefty $150 million Series A3 funding round to scale up its platform in the intelligent manufacturing, entertainment and robotics industries.

The startup attracted a broad base of investors, including the automotive-focused venture capital firm Geely Capital, and video games developers such as 4399 Network, Tanwan and Giant Network. Other backers include Fosun Capital, Orinno Capital, CoStone Capital, Addor Capital, T-Capital, and Muhua Tech Ventures.

Tripo AI, officially known as Holymolly Ltd., develops 3D artificial intelligence models that enable the creation of interactive content for games developers and media creators. The development of high-quality 3D assets in video games, manufacturing and movie production has always been extremely resource-intensive, requiring hundreds of hours of manual labor from skilled digital artists. But with Tripo AI, developers can automate this process, enabling anyone to generate high-fidelity 3D objects using natural language prompts that describe how they want them to look.

The company has seen rapid momentum over the last six months, debuting its flagship 3D foundation models Tripo H3.1 and Tripo P1.0, which introduced significant algorithmic breakthroughs in areas such as 8K resolution texture generation and object segmentation. Those updates were key, because generative 3D environments have long been plagued by low-resolution textures. With its higher-fidelity models, Tripo AI can generate much more realistic simulations that meet the quality standards of modern video games and media content.

Tripo AI is also a developer of world models, or AI-generated 3D environments that accurately reflect real-world physics. Its most recent development here is Project Eden, which is a world model research preview featuring a native architecture that’s able to decouple the underlying state simulation from visual rendering. It’s an entirely new approach that paves the way for “embodied intelligence” applications such as robotics training.

Following today’s round, Tripo AI says its goal is to move beyond cutting-edge research and build up a global commercial ecosystem, but doing so won’t be easy. To get there, it said it needs to invest in three key areas, including core algorithm development to enhance the performance, speed and accuracy of its 3D generation capabilities. It will also build out its data infrastructure to train even larger and more sophisticated models, while striving to recruit “top-tier AI researchers and engineers” that can help it to maintain its competitive lead in the world model sector.

Image: Tripo AI