Investors are betting on the potential of generative AI and making Stability AI, the startup behind Stable Diffusion, a unicorn.
With Stable Diffusion, London-based startup Stability AI released an open-source alternative to DALL-E 2 and co. this summer. Now the company is announcing a $101 million funding round at a $1 billion valuation. The money comes from venture capitalists Coatue, Lightspeed Venture Partners and O'Shaughnessy Ventures.
Stable diffusion is at least on par with DALL-E 2 in terms of usage
The funding is remarkable given the turbulent economic times with rising interest rates that make it difficult for risky startups with uncertain business models to raise large sums of money. However, the potential of generative AI is seen as great by some investors such as leading Silicon Valley venture capitalist Sequoia.
According to Stability AI, the open-source version of Stable Diffusion has been downloaded and licensed 200,000 times by developers since its release. The commercially oriented web version, Dreamstudio, has well over one million users from 50 countries who have generated more than 170 million images to date, according to Stability AI.
Across all channels, Stable Diffusion has more than ten million daily active users, says Stability AI CEO Emad Mostaque. Competitor OpenAI announced more than one million users for DALL-E 2 in early September 2022.
Stability AI plans to bring more generative AI to market
Stability AI plans to use the new funding to further expand its generative offering: Plans include AI models for images, voice, audio, video, and 3D content for consumers and businesses. In addition to Stable Diffusion, Stability AI is also involved in the development of the generative audio model Dance Diffusion.
"There’s been this promise of interactive personalized content for many years. This is the first technology that can deliver it," Mostaque says.
The startup plans to stick to its approach of putting its models into the hands of developers as open source. Instead, Stability AI wants to make money with privately trained models, with infrastructure services, and via the Dreamstudio web platform and API.