Runway raised $315 million in a Series E round led by General Atlantic, which brings its total funding since 2018 to $860 million. The company holds a $5.3 billion valuation, an increase from its $3.3 billion valuation during its April round. Investors including Nvidia, Adobe Ventures, AMD Ventures, Fidelity Management & Research Co., and Felicis Ventures participated in this financing. Runway is a New York-based company that builds video generation models like Gen-4. I find Gen-4 to be a tool for directors who want to control camera movement and character consistency rather than just generating pixels. The company describes itself as an applied AI research company that builds world models for universal simulation. Global funding for AI-related video companies in 2025 reached $3.08 billion, a 94.6% increase from 2024. This growth is evident as companies like Luma AI announced a $900 million Series C at a $4 billion valuation. Runway is growing extremely fast according to Michelle Kwon, the head of operations and partnerships.
Directorial Control in Gen-4
The Gen-4 model includes enhanced text prompt comprehension for intricate designs and provides the Motion Brush to paint motion onto specific areas of a static image. It also includes Act-One to transfer motion from a source video to a generated character. This toolset provides functionality for the director, the VFX artist, and the 3D generalist. I find the camera data export more useful than the generation quality for professional VFX workflows. The model exports camera tracking data as JSON or FBX files for Blender, Cinema 4D, or Adobe After Effects. When you export the FBX camera data from a Gen-4 generation into professional software like Blender or After Effects, the 3D camera in your scene moves in perfect synchronization with the camera from the AI-generated footage. You can import this data to sync 3D elements with the movement of the AI shot. You might find the 1080p resolution a limitation for 4K projects. The model lacks native alpha channel support, so you must use background removal tools to composite elements like smoke or fire.
| Plan | Monthly Price | Credits | Best Use |
|---|---|---|---|
| Free | $0 | 125 (one-time) | Interface testing |
| Standard | $15 | 625 | Occasional work |
| Pro | $35 | 2,250 | Regular creators |
| Max | $95 | 9,500 | Heavy production |
| Enterprise | Custom | Custom | Large companies |
Runway uses a credit-based subscription model that suits iterative creative work. A failed generation costs a few credits instead of a set dollar amount per second like Google Veo 3, which charges $0.75 per second.
Enterprise Security and Licensing
Lionsgate maintains an exclusive deal to use Runway to train a model customized to its proprietary film and television content. This partnership helps filmmakers use AI for pre-production and post-production. Enterprise users receive different terms than standard subscribers. Runway does not use enterprise content for training and provides uncapped IP indemnification. You own all rights to your inputs and outputs under an Enterprise agreement. Runway also maintains SOC 2 Type II certification and ISO/IEC 27001. Enterprise customers can request SOC 2 Type II reports through the sales team. These customers also get API access and custom capacity above the self-serve tiers. Runway supports zero data retention and custom deletion windows. I suggest you check if your specific workflow requires the custom capacity of an Enterprise plan. Does the loss of character consistency in some models undermine the utility of these large-scale training deals?
