SpAItial

SpAItial is a London-based AI company established in 2024 by a team of technology industry veterans: Matthias Niessner, Luke Rogers, Ricardo Martin-Brualla, and David Novotny. Niessner, a co-founder of the AI video generation company Synthesia and a professor at TU Munich, brings deep academic and entrepreneurial experience in AI. The founding team also includes Luke Rogers, formerly of McKinsey and an early leader at Cazoo; Ricardo Martin-Brualla, who led generative 3D initiatives at Google; and David Novotny, who headed 3D asset generation at Meta.

The company is developing Spatial Foundation Models (SFMs), an AI system designed to understand, generate, and reason about 3D environments with an intrinsic comprehension of space-time and physics. Unlike generative AI models that operate on pixels or text, SpAItial's SFMs work natively with 3D structures, capturing geometry, material properties, and physics from minimal inputs like a single image, a short video, or text. This technology aims to produce photorealistic and physically consistent virtual worlds. The company's mission is to create AI that can seamlessly navigate and connect virtual and physical realms.

SpAItial launched from stealth mode with a $13 million seed funding round led by Earlybird Venture Capital, with participation from Speedinvest and a number of angel investors. The company's business model is expected to involve licensing its foundation model to developers. The technology has broad potential applications across various industries, including creating immersive worlds for gaming and entertainment, advancing CAD engineering and construction, powering VR/AR experiences, and developing physically-intelligent robotics. Early demonstrations have shown the model's capability to generate photorealistic 3D environments from a single image.

Keywords: Spatial Foundation Models, generative 3D, 3D world generation, AI, spatial intelligence, physics-based AI, virtual environments, augmented reality, virtual reality, robotics, digital twinning, 3D content creation, immersive entertainment, CAD engineering, AI-native 3D, spatio-temporal AI, Matthias Niessner, generative AI, deep tech, computer vision