Stability AI, the company that funds the development of open source music and image-generating systems such as Dance Diffusion and Stable Diffusion, today announced it has raised $101 million in a round of funding led by Coatue and Lightspeed Venture Partners with the participation of O’ Shaughnessy Ventures LLC. The tranche values the company at $1 billion post-money, according to to a Bloomberg resource, and comes as demand for AI-powered content generation accelerates.
Stability AI in London and San Francisco is the brainchild of CEO Emad Mostaque. After graduating from Oxford with a master’s degree in mathematics and computer science, he served as an analyst at several hedge funds before moving into more public works. Mostque co-founded Stability AI in 2020, motivated by both a personal fascination with AI and what he described as a lack of “organization” within the open source AI community.
“Nobody has the right to vote except our employees — not billionaires, big funds, governments or anyone else with control over the company or the communities we support. We are completely independent,” Mostaque told londonbusinessblog.com in an earlier interview. “We plan to use our computing power to accelerate open source, fundamental AI.”
Stability AI has a cluster of more than 4,000 Nvidia A100 GPUs running in AWS, which it uses to train AI systems, including Stable Diffusion. It’s Quite Expensive to Maintain — Business Insider reports that Stability AI’s operations and cloud spend exceeded $50 million. But Mostaque has repeatedly claimed that the company’s R&D will enable it to train models more efficiently in the future.
The company’s work has not escaped controversy, it is worth noting. The open source release of Stable Diffusion has been used to create offensive content such as graphic violence and pornographic, non-consensual celebrity deepfakes. And Getty Images banned uploading content generated by systems like Stable Diffusion, for fear of intellectual property disputes. (Stable Diffusion is trained on a dataset that contains copyrighted works — and even personal medical records.)
In fact, stability AI was the subject of a critical recent letter from U.S. House Representative Anna G. Eshoo (D-CA) to the National Security Advisor (NSA) and the Office of Science and Technology Policy, urging her to join the NSA and OSTP. urged to address the release of “unsafe AI models” that “do not moderate content created on their platforms”.
Stability AI has taken a largely hands-off approach to moderation to date, including filtering tools with the open source Stable Diffusion package, but allowing users and businesses to deploy the system however they want, as long as they adhere to the terms. of the company’s license. “A percentage of people are just unpleasant and weird, but that’s human,” Mostaque said in an earlier interview. “Indeed, it is our belief that this technology will be widespread, and the paternalistic and somewhat condescending attitude of many AI enthusiasts is misleading because they do not trust society.”
Stability AI plans to monetize by training “private” models for customers and acting as a common infrastructure layer. It also provides a platform and API, DreamStudio, which allows individual users to access the models. Mostaque told Bloomberg that DreamStudio has more than 1.5 million users who have created more than 170 million images and that Stable Diffusion has more than 10 million daily users. “on all channels.”
The open source version of Stable Diffusion has since been downloaded more than 200,000 times, according to a press release published this morning by Stability AI.
According to Mostaque, the capital from the funding round will support the wider deployment of customized versions of Stable Diffusion for users and investing in more supercomputing power. It will also focus on hiring more people, with Mostaque saying he expects to grow from 100 to about 300 employees over the next year.
Stability AI has recently made several high-profile hires, attracting research scientists from Google Brain and futurist and public speaker Daniel Jeffries.
In addition to Stable Diffusion, Stability AI claims to have other commercializable projects in the works, including AI models for generating audio, language, “3D” and even video. One is the aforementioned Dance Diffusion, an algorithm and set of tools that can generate music clips by training hundreds of hours of existing songs.
Sri Viswanath, a general partner at Coatue, said in a statement: “At Coatue, we believe that open source AI technologies have the power to unlock human creativity and achieve a broader good. Stability AI is a big idea that dreams further beyond the immediate applications of AI. We are excited to be a part of Stability AI’s journey and look forward to seeing what the world creates with Stability AI technology.”