Stability AI CEO Resigns, Citing Inability to Surpass Centralized AI with a Similar Approach

Stability AI founder and chief executive Emad Mostaque has stepped down from the top role and the unicorn startup’s board, the buzzy firm said Friday night, making it the second hot AI startup to go through major changes this week.

Stability AI, which has been backed by investors including Lightspeed Venture Partners and Coatue Management, doesn’t have an immediate permanent replacement for the CEO role but has appointed its COO Shan Shan Wong and CTO Christian Laforte as interim co-CEOs, it said in a blog post.

Stability AI, which has lost more than half a dozen key talent in recent quarters, said Mostaque is stepping down to pursue decentralized AI. In a series of posts on X, Mostaque opined that one can’t beat “centralized AI” with more “centralized AI,” referring to the ownership structure of top AI startups such as OpenAI and Anthropic.

He additionally asserted that it was his decision to step down from the top role as he held the most number of controlling shares. “We should have more transparent & distributed governance in AI as it becomes more and more important. Its [sic] a hard problem, but I think we can fix it..,” he added. “The concentration of power in AI is bad for us all. I decided to step down to fix this at Stability & elsewhere.”

Mostaque’s departure from Stability AI, a startup known for its popular image generation tool Stable Diffusion, comes amid an ongoing struggle at the startup that was spending a reported estimate of $8 million a month as of October 2023, according to Bloomberg, which also noted that the startup had unsuccessfully attempted to raise new funding at a $4 billion valuation.

Mostaque, it appears, wasn’t prioritizing revenue growth about a year ago. In a post on X last year, he expressed his amusement at the generative AI companies’ “strange focus on revenue” even as “the technology is useful but far from vaguely mature as new breakthroughs happen almost daily.” He cited several examples, including MagicLeap, which spent billions before generating revenue.

“The payoffs on proper generative AI R&D are clearer and faster to market than just about anything we’ve seen. It’s going to create way more economic value than self-driving cars, for example; the total investment in that has been $100b with no revenue payoff,” he wrote.