Synthesia cofounder Matthias Niessner has raised $13m for his newly launched startup, TechCrunch reported on Tuesday.
The Munich-based startup, named Spaitial, aims to produce models capable of generating fully interactive, photorealistic 3D environments from text prompts. This would mark a significant leap forward from current capabilities like OpenAI’s DALL-E or Stable Diffusion which focus on 2D images.
The seed round was led by Earlybird Venture Capital, an early stage investor that backed UiPath through its Digital East Fund, with participation from Speedinvest and several high profile angels like Black Forest Labs CEO Robin Rombach and Synthesia CEO Victor Riparbelli.
Despite having only released a teaser video demonstrating text-to-3D room generation, the company has already garnered significant interest.
The team includes Ricardo Martin-Brualla, formerly of Google’s 3D teleconferencing platform Beam, and David Novotny, who led Meta’s text-to-3D asset generation project.
Niessner, a professor at the Technical University of Munich, is widely regarded as one of Europe’s leading researchers in 3D AI modelling. He cofounded Synthesia, the AI avatar startup valued at $2.1bn, in London in 2017 with three fellow AI researchers from UCL, Stanford and Cambridge.
He told TechCrunch his vision goes beyond creating static 3D worlds — he aims to ensure they are interactive and behave like real environments, a challenge he believes remains largely unresolved in the field.
The startup is up against competitors such as San Francisco-based Odyssey, a developer of AI video generation tools for the film industry, and World Labs founded by renowned Stanford University AI professor Dei-Dei Li which is already valued at over $1bn.
