Hard Fork AI Runway Raises $315M for AI World Models
Feb 11, 2026
A deep dive into Runway's $315M raise and its push to pretrain next‑gen world models for video. Explanations of world models, 3D simulation and why they matter for realistic camera moves. Coverage of Gen 4.5 capabilities and how physics-aware tools reduce glitches. Discussion of competing labs, compute deals, and the evolving regulatory landscape around deepfakes and safety.
AI Snips
Chapters
Transcript
Episode notes
World Models Enable Coherent Video
- Runway is investing its $315M to pre-train 'world models' that simulate full 3D environments for video AI.
- These models let the AI reason about space, physics, and camera movement to produce coherent multi-angle footage.
3D Simulation Prevents Glitchy Videos
- Video generation needs simulated 3D environments so camera pans and angle changes remain consistent.
- Without spatial understanding, generated videos become glitchy when the viewpoint changes.
Runway's Start With Physics-Aware Video
- Runway first gained traction with physics-aware video tools favored by media and advertising.
- The company then released its first world model and sees that technology as foundational beyond just creative work.