Stability AI
Stability AI, the company behind the innovative text-to-image AI model known as Stable Diffusion, has unveiled a groundbreaking AI application, Stable 3D. This software is set to revolutionize 3D model creation for the world of modeling and game development, catering to platforms such as Blender, Maya, Unreal Engine, and Unity.
Currently, Stable 3D is available for a select group of customers who can request access through Stability’s contact form. The key goal of Stable 3D is to empower non-experts, enabling them to produce “draft-quality” 3D models within minutes, according to Stability AI’s blog.
The creation of 3D content has traditionally been a complex and time-consuming task, particularly for graphic designers, digital artists, and game developers. It often takes hours, if not days, to craft a moderately intricate 3D object. However, Stable 3D aims to level the playing field by allowing independent designers and developers to generate thousands of 3D objects daily at minimal cost.
Stable 3D offers robust capabilities, comparable to other model-generating tools available. Users can describe the 3D model they want to create in natural language or upload existing images or illustrations for conversion into 3D models. These models are generated in the “.obj” file format, making them compatible with most standard 3D modeling tools for further editing and manipulation.
Notably, Stability AI has not disclosed the data used to train Stable 3D. Given the risk associated with generative AI models potentially reproducing copyrighted training data, this could raise concerns among the tool’s commercial users. Unauthorized use of copyrighted material could result in legal issues for customers incorporating IP-infringing work into their projects.
Stability AI has faced IP-related challenges in the past, with Getty and various artists suing the company for allegedly using their images without proper notification or compensation to train Stable Diffusion. Although Stability AI has recently partnered with startup Spawning to address “opt-out” requests from artists, it remains unclear if this partnership extends to Stable 3D’s training data.
Despite these potential legal concerns, Stable 3D signifies Stability AI’s entry into the competitive field of AI-powered 3D model generation, which is already teeming with players. Notable competitors include 3D object-creating platforms like 3DFY and Scenario, as well as startups such as Kaedim, Auctoria, Mirage, Luma, and Hypothetic. Established companies like Autodesk and Nvidia have also ventured into this space with apps like Get3D and ClipForge.
Moreover, tech giants like Meta and OpenAI have explored methods to generate 3D assets from prompts, showing the growing interest in this domain. Stable 3D is perhaps Stability AI’s latest effort to diversify its business and stay competitive in the face of increasing competition, including from art-creating generative AI platforms like Midjourney and OpenAI.
Recent reports have indicated financial challenges for Stability AI, with cash burn and payroll issues. Despite raising $25 million through a convertible note, the company has yet to secure new funding at a higher valuation, maintaining a valuation of $1 billion. Nevertheless, Stability AI continues to introduce new features in its portfolio, such as personalized model fine-tuning and a “sky replacer” tool in its AI-powered photo editing suite.
