Archive — US — techcrunch
Stability AI’s new AI model turns photos into 3D scenes
Stability AI has introduced Stable Virtual Camera, an AI model that transforms 2D images into immersive 3D videos. It allows users to generate novel views from images, controlling camera angles and paths. The model is currently in research preview, available for non-commercial use, but may produce lower-quality results in complex scenes. The firm faces recent management changes and seeks to recover from past challenges.
Key points
- Stability AI launched Stable Virtual Camera, an AI tool for converting 2D images into immersive videos.
- The model generates videos with customizable camera angles and dynamic pathways from up to 32 input images.
- It supports various aspect ratios (1:1, 9:16, 16:9) with a max length of 1,000 frames.
- Warnings include potential low-quality output in scenes with humans, animals, or complex textures.
- Stability AI has faced financial difficulties and management issues, leading to recent leadership changes.
Why it matters
- The trend of using generative AI in multimedia is growing, impacting industries like gaming and film.
- The introduction of AI tools for video creation aligns with advancements in machine learning and design ecosystems as seen in 2025 IT trends.
- Stability AI's efforts reflect broader challenges in the startup ecosystem, highlighting the importance of strong leadership and investor confidence.
Stability AI's new model represents a significant advancement in generative technology, providing innovative tools for content creation while navigating challenges typical of emerging AI firms.