World Labs 獲 Autodesk 2 億美元，將世界模型融入 3D

• World Labs develops multimodal world models capable of understanding and generating realistic, persistent 3D environments with spatial reasoning[3] • Models are trained to comprehend and reason about geometry, physics, dynamics, and semantic relationships within 3D spaces[2][3] • Autodesk's 'neural CAD' represents a complementary approach: generative AI trained on geometric data that can reason about components and entire systems[1] • Neural CAD generates working 3D models (not just images) with understanding of real-world functionality and design constraints[1] • The partnership enables research and model-level collaboration to explore how world models can integrate with Autodesk's existing geometry and simulation software infrastructure[3] • Initial focus on entertainment use cases, leveraging Autodesk's established workflows in design and construction sectors[1]

This investment signals a strategic shift in AI development toward spatial intelligence and physical-world reasoning. The partnership positions Autodesk to embed advanced spatial AI across its design and construction software portfolio, potentially transforming how professionals approach 3D modeling and simulation. By combining World Labs' world model capabilities with Autodesk's four decades of geometry and simulation expertise, the collaboration could establish new standards for AI-assisted design workflows. The emphasis on 'physical AI' that understands real-world constraints—geometry, physics, materials—addresses a critical gap in current AI systems and could unlock applications across architecture, engineering, manufacturing, and entertainment. The reported $5 billion valuation trajectory suggests investor confidence that spatial AI represents a major frontier comparable to or distinct from large language models, potentially creating a new category of enterprise AI tools.