OpenAI co-founder Sutskever’s Startup Set to Raise $1 Billion: Reports

Ilya Sutskever, co-founder at OpenAI, is reportedly raising over $1 billion for his startup Safe Superintelligence Inc (SSI), which is valued at more than $30 billion, as reported by Bloomberg. 

The deal is being led by San Francisco-based venture capital firm Greenoaks Capital Partners, which is set to invest $500 million, according to the report. Greenoaks has previously invested in AI companies like Scale AI and Databricks.

In September last year, Sutskever announced that the lab had raised $1 billion from NFDG, a16z, Sequoia Capital, DST Global, and SV Angel.

Inside SSI’s Mission

The company, based in Palo Alto with offices in Tel Aviv, was launched in June last year. SSI is led by Sutskever, entrepreneur and investor Daniel Gross, and former OpenAI employee Daniel Levy. Gross previously co-founded AI startup Cue, which Apple acquired in 2013 for between $40 million and $60 million.

SSI has established the world’s first lab dedicated solely to developing safe superintelligence. The company’s mission is clear: to build a safe superintelligence. 

“We will pursue safe superintelligence in a straight shot, with one focus, one goal, and one product. We will do it through revolutionary breakthroughs produced by a small cracked team,” Sutskevar said. 

The company emphasised that safety and capabilities will be addressed simultaneously, as technical problems require revolutionary engineering and scientific breakthroughs. SSI aims to rapidly advance capabilities while ensuring that safety remains paramount. 

Ilya’s Vision

Discussing the unpredictable nature of AI reasoning at the NeurIPS 2024, Sutskever said as AI systems reason more, they become less predictable. He highlighted how advanced AI, like AlphaZero, already demonstrates unpredictability in areas such as chess. Sutskever predicted that artificial superintelligence will evolve to understand complex concepts with limited data and self-correct, reducing errors like hallucinations. He also noted that future AI systems will become truly agentic.