OpenAI Co-Founder Launches Safe Superintelligence Startup Focused on Ethical AI Development

Ilya Sutskever formally launched Safe Superintelligence Inc. this week, the startup he has been building since leaving OpenAI in May 2024. The company's premise is deliberately narrow: build a safe artificial general intelligence and nothing else. No products. No API. No enterprise sales. Just a pure research organization devoted to ensuring that the first AGI does not destroy the people who built it. The company has raised funding from investors including Sequoia Capital and Andreessen Horowitz, though the amount has not been disclosed.

Sutskever's exit from OpenAI was one of the most significant personnel moves of 2024. As chief scientist, he had been central to the development of GPT-3, GPT-4, and the research direction that produced ChatGPT. He was also one of the board members who voted to fire Sam Altman in November 2023, only to reverse his position days later and sign a letter demanding Altman's reinstatement. The episode left his position at the company untenable. In May, he announced his departure, and in June, he revealed SSI.

"We have started the world's first straight-shot SSI lab, with one goal and one product: a safe superintelligence. We approach safety and capabilities in tandem, as technical problems to be solved through revolutionary engineering and scientific breakthroughs." – Ilya Sutskever

The launch puts SSI in an unusual position within the AI landscape. OpenAI, Google DeepMind, and Anthropic all pursue safety research alongside commercial products that generate billions in revenue. Sutskever's bet is that the tension between safety and commercialization is unresolvable within a single organization. If you are racing to ship a product that competes with GPT-5 and Gemini, he appears to be arguing, you will cut corners on the safety research that might tell you not to ship it at all.

The question hanging over SSI is whether a pure research organization can attract and retain the talent needed to achieve AGI when competitors offer equity packages worth millions. Sutskever has answered that by recruiting Daniel Gross, a former Apple AI executive, and Daniel Levy, a former OpenAI researcher. The company's policy of not disclosing its funding makes it impossible to assess whether SSI has the resources to compete with the hyperscalers, but the caliber of its early hires suggests it is not undercapitalized.