Stability AI, the startup behind Stable Diffusion, raises $101M

Stability AI, the company funding the development of open source music- and image-generating systems like Dance Diffusion and Stable Diffusion, today announced that it raised $101 million in a funding round led by Coatue and Lightspeed Venture Partners with participation from O’Shaughnessy Ventures LLC. The tranche values the company at $1 billion post-money, according to a Bloomberg source, and comes as the demand for AI-powered content generation accelerates.

London- and San Francisco–based Stability AI is the brainchild of CEO Emad Mostaque. Having graduated from Oxford with a master’s in mathematics and computer science, he served as an analyst at various hedge funds before shifting gears to more public-facing works. Mostaque co-founded and bootstrapped Stability AI in 2020, motivated both by a personal fascination with AI and what he characterized as a lack of “organization” within the open source AI community.

“Nobody has any voting rights except our employees — no billionaires, big funds, governments or anyone else with control of the company or the communities we support. We’re completely independent,” Mostaque told TechCrunch in a previous interview. “We plan to use our compute to accelerate open source, foundational 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 costly to maintain — Business Insider reports that Stability AI’s operations and cloud expenditures exceeded $50 million. But Mostaque has repeatedly asserted that the company’s R&D will enable it to train models more efficiently going forward.

The company’s work hasn’t escaped controversy, it’s worth noting. The open source release of Stable Diffusion has been used to create objectionable content like graphic violence and pornographic, nonconsensual celebrity deepfakes. And Getty Images banned the upload of content generated by systems like Stable Diffusion, over fears of intellectual property disputes. (Stable Diffusion was trained on a data set that includes copyrighted works — and even private medical records.)

Stability AI was even the subject of a recent critical 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, in which she urged the NSA and OSTP to address the release of “unsafe AI models” that “do not moderate content made on their platforms.”

Stable Diffusion

A collage of images created by Stable Diffusion. Image Credits: Daniel Jeffries

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 companies to deploy the system how they wish so long as they follow the terms of the company’s license. “A percentage of people are simply unpleasant and weird, but that’s humanity,” Mostaque said in a previous interview. “Indeed, it is our belief this technology will be prevalent, and the paternalistic and somewhat condescending attitude of many AI aficionados is misguided in not trusting society.”

Stability AI plans to make money by training “private” models for customers and acting as a general infrastructure layer. It also offers a platform and API, DreamStudio, through which its models can be accessed by individual users — Mostaque told Bloomberg that DreamStudio has more than 1.5 million users who’ve created over 170 million images and Stable Diffusion has more than 10 million daily users “across all channels.”

Meanwhile, the open source version of Stable Diffusion has been downloaded more than 200,000 times, according to a press release published by Stability AI this morning.

According to Mostaque, the capital from the funding round will support deploying custom versions of Stable Diffusion for users at a larger scale and investing in more supercomputing power. It’ll also be put toward hiring more people, with Mostaque saying he expects to grow to about 300 employees from 100 over the next year.

Stability AI has made several high-profile hires recently, bringing on research scientists from Google Brain and futurist and public speaker Daniel Jeffries.

Beyond 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 of those is the aforementioned Dance Diffusion, an algorithm and set of tools that can generate clips of music by training on 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 beyond the immediate applications of AI. We are excited to be part of Stability AI’s journey, and we look forward to seeing what the world creates with Stability AI’s technology.”