Matthias Niessner, one of Europe’s most respected AI researchers and a co-founder of Synthesia, has launched a new startup named SpAItial with the ambitious goal of transforming text prompts into fully interactive, photorealistic 3D environments.
The Munich-based company has raised a $13 million seed round led by Earlybird Venture Capital, with Speedinvest and prominent angels also participating.
Chasing the ‘holy grail’ of generative models
Unlike today’s image generators like OpenAI’s DALL·E or Stability AI’s Stable Diffusion, SpAItial wants to build foundation models capable of generating not just static 3D visuals, but dynamic, immersive, and physically accurate digital environments — what Niessner calls “worlds that behave like the real world.”
The team behind the vision includes ex-Google and Meta engineers: Ricardo Martin-Brualla, previously with Google’s 3D teleconferencing project Beam, and David Novotny, who led Meta’s text-to-3D initiatives.
Advertisement
A startup built on deep technical roots
SpAItial is still in stealth with little public output besides a teaser demo showing how a text prompt generates a realistic 3D room. But with Niessner’s track record and the strength of his cofounders, investors are betting big on the promise.
The startup’s business model will revolve around licensing its core models to developers for specific applications — ranging from video game development and entertainment to use cases in digital twins and construction.
Prioritizing realism, interactivity, and revenue
While many AI startups emphasize scale and hype, SpAItial is taking a pragmatic approach. The company is intentionally hiring a small, elite team and focusing on creating models that don’t just visualize — they simulate. The aim? Make it possible for a 10-year-old to create a fully playable video game from a text prompt in just 10 minutes.
In Niessner’s words, “It’s not just about generating 3D assets. We want those objects and environments to respond like the real world — to shatter, move, interact.”
A race toward the next frontier in AI
SpAItial enters a competitive space already populated by players like Odyssey and World Labs (founded by Fei-Fei Li), but sees vast white space compared to image generation. Its challenge now will be proving technical feasibility, establishing early partnerships, and unlocking commercial traction — all while aiming to deliver one of AI’s most coveted frontiers: the generative, interactive 3D internet.
