Stability AI Gains New Leadership as Generation AI Innovations Continue to Roll Out

A few months after the founder and former CEO of Stability AI left under a cloud of doubt, the generative AI startup now has a new full-time CEO.

According to a report from Information, Prem Akkaraju, former CEO of visual effects company Weta Digital, is set to become the new CEO of Stability AI. Akkaraju was CEO of Weta Digital since January 2020 and is also co-founder and executive chairman of Projection room. Stability AI declined to provide any official comment on the report to VentureBeat.

The news comes nearly three months after the founder and former CEO of Stability AI. Emad Mostaque resigned in March. Since then, the company has had Stability’s Chief Operating Officer (COO) Shan Shan Wong and Chief Technology Officer (CTO) Christian Laforte serve as interim co-CEOs.

Among the reasons for Mostaque’s departure were investors’ concerns about the financial viability of the company and its business operations. Mostaque founded Stability AI in 2019 to help advance the state of AI. Stability AI is perhaps best known for its stable streaming text-to-image generative AI technology, which the company has continued to improve. On June 12, Stable diffusion medium 3 was released as the fastest and most capable version of the company’s text-to-image generation templates.

A Reuters report published on May 16, claimed that Stability AI lost more than $30 million in the first quarter of the year, on revenue of less than five million. That same report claimed that Stability AI was seeking outside investment. The report of Information also claims that alongside the new CEO, Stability AI will receive a cash injection from a group of investors led by Sean Parker.

Stability AI’s growing set of models and activities

Stability AI has faced increasing competitive pressure over the past year. In the image generation space where Stable Diffusion exists, Dall-E 3, Midjourney, Ideogram and others from OpenAI are increasingly commoditizing the market.

Stability AI has branched out beyond just generating images, however. The company is also developing a code model called Stable code which had a big release at the start of the year. Stability AI also has a large language model called Stable ML which has multiple model sizes, including 1.6 billion, 3 billion, and 7 billion parameter models.

Audio and video are also areas where Stability AI is active. Stable Audio 2.0 was released on April 3, allowing users to generate up to 3 minutes of audio. Stable Video provides a set of text to video generation, including the ability to do 3D with Stable 3D video (SV3D).

Underpinning all of these AI models is an updated business model that debuted late last year. In December 2023, Stability AI announced a membership model with the basic idea that organizations that want to use the technology for commercial purposes must pay a membership. Stability AI continues to make some weights from some of its models available under a non-commercial license.

Beyond membership, the company offers access via an API, where organizations pay based on usage. This API can be accessed directly or via an interface such as Stable craftsman Discord bot released on May 9.

Stability AI founder working on sequel

Although Stability AI now has a new CEO, the old CEO is still busy.

Mostaque posted publicly on X (formerly Twitter) that he was working on a new company known as Schelling AI. Not much is known about Schelling AI at the moment, other than that it aims to help advance Mostaque’s vision for decentralized AI.

Mostaque told VentureBeat that more details about what Schelling AI actually does will be revealed in July.