Rapid Media
EN ไทย
AI Can Read Everything in the World, But Doesn't Know If a Glass Will Break When Dropped—As the World Moves from LLMs to World Models, Where Will Investment Flow?
78 Techsauce · Aug 10, 2026
AI Can Read Everything in the World, But Doesn't Know If a Glass Will Break When Dropped—As the World Moves from LLMs to World Models, Where Will Investment Flow?
Technology
▾ Hover for summary ▾ Tap for summary
AI Can Read Everything in the World, But Doesn't Know If a Glass Will Break When Dropped—As the World Moves from LLMs to World Models, Where Will Investment Flow? Techsauce · Aug 10, 2026

The AI industry is shifting from Large Language Models (LLMs) to World Models—systems that understand spatial intelligence, physics, and 3D reasoning. Within 60 days in early 2026, World Labs and AMI Labs closed over $2 billion in funding to build AI systems that grasp concepts like gravity and object dynamics. Two competing approaches are emerging: "Renderers" (visual synthesis via Marble, Genie 3) and "World Understanders" (physics-accurate prediction via Cosmos, JEPA). This shift signals a move from chatbot interfaces toward AI that can operate in physical environments, drawing massive VC capital flows from VR/AR and beyond.
Read at Techsauce →

← Back to all articles