OpenAI co-founder and former chief scientist Ilya Sutskever has announced the launch of a new company, Safe Superintelligence Inc. (SSI). The venture comes shortly after Sutskever’s departure from OpenAI, marking a significant shift in his career towards a sole focus on AI safety. SSI, which Sutskever founded with former Apple AI lead Daniel Gross and former OpenAI engineer Daniel Levy, aims to develop a safe and powerful AI system by prioritizing capabilities as well as safety.

Announcing his new company on X, Sutskever wrote, “I’m starting a new company.” In another tweet, he explained that his new company will pursue “secure superintelligence in a straight shot, with the same mission, the same goals, and the same product.”

SSI’s announcement highlights the company’s commitment to advancing AI technology, while making sure to keep security at the forefront. Sutskever’s vision for SSI is clear: “one goal and one product,” focusing specifically on creating super intelligent AI systems that are both advanced and safe. This approach allows SSI to avoid the distractions and pressures often faced by larger AI firms such as OpenAI, Google, and Microsoft, which have to balance innovation with commercial and management demands.

Sutskever’s departure from OpenAI earlier this year made a lot of headlines. He played a key role in the effort to oust OpenAI CEO Sam Altman, which led to a period of internal conflict. After this tumultuous period, Sutskever expressed regret for his involvement, and emphasized his dedication to the mission they built together at OpenAI. This experience seems to have shaped his approach at SSI, where his focus is on maintaining a steady and uninterrupted path towards safe AI development.

SSI’s approach to AI safety is based on lessons learned from OpenAI. At OpenAI, Sutskever co-led the Superalignment team with Jane Leakey, who left the company in May to join Anthropic, a rival AI firm. The Superalignment team was dedicated to governing and running AI systems to ensure they remain beneficial. This mission continues at SSI, where safety and capabilities are treated as interconnected challenges to be addressed through innovative engineering and scientific breakthroughs.

In an interview with Bloomberg, Sutskever outlined SSI’s business model, which is designed to keep safety, security, and progress separate from short-term commercial pressures. This structure allows the company to focus entirely on its mission without the distractions of management overhead or product cycles. Unlike OpenAI, which evolved from a nonprofit to a for-profit entity due to the high costs of AI development, SSI has been set up as a for-profit company from the start, with a clear emphasis on raising capital to support its ambitious goals.

SSI is currently building its team and has established offices in Palo Alto, California and Tel Aviv. The company is actively recruiting technical talent to join its mission to create secure superintelligence.