Creating Interactive 3D Worlds from Text Just Got M

Generative AI has proven to be quite powerful. Today, you can generate hyper-realistic images, transform pictures by applying multiple styles, animate static photos, or get impressive video clips of practically anything you want, all with a simple prompt. However, things could go beyond “simple” visual or auditory content. A leading European AI researcher pursuing an AI tool capable of generating interactive content has just secured a significant $13 million in funding.

Forget static images and flat videos from AI. Imagine typing a few words and instantly conjuring up fully interactive, photorealistic 3D environments that behave just like the real world. This ambitious vision is precisely what Matthias Niessner, a prominent name in AI 3D modeling, is chasing. Niessner launched “SpAItial,” his new startup, to make it a reality.

SpAItial works on AI capable of generating interactive 3D environments from text

The company’s mission is nothing short of cracking what some are calling the “holy grail” of AI models. In other words, AI that can generate complete, coherent, and interactive 3D worlds from simple text prompts. Think text-to-image, but for entire virtual environments.

We’ve already seen incredible strides in AI’s ability to create stunning 2D visuals and even short video clips. However, generating dynamic 3D worlds that understand physics and respond to interactions is a whole different ball game. This is where SpAItial aims to make its mark, going beyond just pretty scenes to build worlds where a glass can shatter realistically or a character can navigate naturally.

Brilliant talent behind the startup

The $13 million seed funding round was led by Earlybird Venture Capital. It is a strong vote of confidence in Niessner’s vision and SpAItial’s approach. The team behind SpAItial isn’t just brilliant; it’s packed with talent, including experts like Ricardo Martin-Brualla (formerly of Google 3D) and David Novotny (from Meta’s text-to-3D efforts). This combined experience gives them a powerful edge in this nascent but highly competitive field.

SpAItial’s ultimate goal is to empower users, even a 10-year-old, to type text and create an entire video game in minutes. But the applications extend far beyond gaming. Imagine the possibilities for digital twins in industrial settings, realistic environments for augmented reality (AR), detailed visualizations for construction projects, and even training grounds for robotics.

The company plans to initially partner with developers, licensing its foundational models for specific use cases. It will also focus on quality over sheer quantity in their hiring strategy. As AI continues to push boundaries, SpAItial’s pursuit of truly interactive, text-generated 3D worlds marks an exciting new frontier, promising to reshape how we create, interact with, and experience digital environments.