London-based SpAItial, a startup focused on how AI understands and generates 3D environments, has emerged from stealth mode with a €11.4 million Seed round in order to scale the development of the company’s first Spatial Foundation Model (SFM).
The round was led by Earlybird Venture Capital, joined by Speedinvest and several angels including: Robin Rombach (Co-Founder & CEO at Black Forest Labs), Victor Riparbelli (Co-Founder & CEO at Synthesia), Steffen Tjerrild (Co-Founder & COO at Synthesia), Christian Stiebner, Edward Grefenstette, Maximilian Odendahl (ex CEO Silexica), Jon Barron, Rahul Garg, Sergej Epp (CISO at Sysdig and Palo Alto Networks), Elias Schneider (CEO Codesphere), Paul Whitehead (CEO Zoopla).
“Current AI models generate images pixel by pixel — great for snapshots, but not for immersive, coherent worlds,” said Matthias Niessner, Co-founder of SpAItial and Professor at TU Munich. “At SpAItial, we’re building AI that’s grounded in space and time from the start. Our spatial foundation models are a game changer for any application that depends on 3D understanding.”
Founded by Matthias Niessner, Luke Rogers, Ricardo Martin-Brualla, and David Novotny in 2024, SpAItial innovates in SFMs, an AI paradigm designed to generate and reason about the appearance and physics of real and imagined environments. SFMs possess an intrinsic understanding of space-time, enabling transformative shifts in applications at the intersection of virtual and physical worlds.
According to SpAItial, their early demo results show their model can generate photorealistic 3D worlds from a single image, with applications ranging from immersive entertainment to robotics and digital twinning.
Unlike today’s models that generate pixel by pixel, SpAItial’s SFM operates directly on 3D structures. This allows it to bridge virtual and physical environments with unprecedented accuracy and realism.
From creating high-fidelity game worlds to enabling robots to better navigate dynamic environments, SpAItial’s technology has the potential to unlock a new class of AI-native 3D applications, with applications in entertainment, urban planning, and industrial automation.
According to comments by SpAItial, industries ranging from the €193 billion gaming sector to robotics and digital twins are increasingly dependent on spatial data to simulate, design, and automate the physical world. However, today’s 3D creation tools remain outdated, slow, and often inaccessible to non-experts. SpAItial aims to change this by building the underlying infrastructure for more accessible 3D generation—streamlining design workflows and enabling new applications in AR/VR, eCommerce, and autonomous systems.
“We couldn’t be more excited to back SpAItial’s team of world-class technologists,” said Dr Andre Retterath, General Partner at Earlybird. “While text, image, and video have seen massive generative AI progress, 3D remains stuck in the pre-AI era. SpAItial’s unique depth in both academic research and real-world product development positions them to bring spatial intelligence into reality.”
With the fresh financing, SpAItial plans to expand the team with talent in AI, graphics, and 3D systems to build foundational technology, and accelerate the development of SpAItial’s first-generation SFMs. They’ll be moving early prototypes into scalable infrastructure and scaling up models with advanced physics capabilities, leading to more commercial pilots across entertainment, digital twinning, robotics, and beyond.
Read the orginal article: https://www.eu-startups.com/2025/05/from-video-games-to-urban-planning-british-spaitial-emerges-from-stealth-with-e11-4-million-to-develop-their-3d-modelling-ai/
