Stability AI Ltd. today appointed Prem Akkaraju chief executive officer and announced that it has raised funding from a group of prominent investors.

The development comes three months after the startup’s founding CEO, Emad Mostaque, stepped down. His resignation was reportedly preceded by several other executive departures. A number of key engineers are believed to have left as well around the same time.

Stability AI develops Stable Diffusion, a series of open-source artificial intelligence models optimized for image generation and editing tasks. The company disclosed today that the models have been downloaded more than 150 million times to date. Alongside Stable Diffusion, Stability AI offers neural networks for generating audio files, videos and code.

Akkaraju (pictured) previously led Academy Award-winning special effects studio Weta Digital. The studio’s software unit was acquired by Unity Technologies Inc. in 2019 for $1.625 billion. Akkaraju is succeeding Shan Shan Wong and Christian Laforte, who took over as Stability AI’s co-CEOs when Mostaque left in March. 

Akkaraju is also a member of the investor group from which Stability raised its newly disclosed funding round. The consortium’s other participants are former Google LLC CEO Eric Schmidt, former Facebook president Sean Parker and venture capitalist Robert Nelsen. They were joined by Greycroft, Coatue Management, Sound Ventures, Lightspeed Venture Partners and O’Shaughnessy Ventures. 

The Wall Street Journal reported that the investment is worth $80 million. Additionally, the consortium reportedly persuaded Stability AI suppliers to forgive more than $100 million in debt and $300 million in future spending obligations.

Forbes previously reported in March that the company’s expenses significantly exceeded its revenue. Last November, Bloomberg cited sources as saying that Stability AI was spending $8 million a month to support its business operations. The AI developer financed its work with a $101 million funding round it had raised in 2022.

Despite its financial challenges, the company has continued investing in product development. Earlier this month, Stability AI released a new open-source AI model called Stable Audio Open. It allows users to create sound effects using natural language prompts. According to Stability AI, Stable Audio Open is capable of generating clips up to 47 seconds in length and can be customized with user-provided datasets.

A week after releasing the model, the company debuted an upgraded iteration of its flagship image generation AI. Stable Diffusion 3 Medium is better at understanding lengthy prompts and can embed text into the images it generates with fewer errors. The model features only 2 billion parameters, which makes it compact enough to run on consumer graphics cards.

“We will continue to release the most cutting-edge open models for the community, while meeting the overwhelming demand for AI solutions from large-scale corporations,” Akkaraju said.

The company plans to build products for not only creative professionals but also developers. Additionally, Stability AI will offer custom AI models tailored to the requirements of enterprise customers.

Photo: Stability AI