Ilya Sutskever, former OpenAI co-founder and now CEO of Safe Superintelligence Inc. (SSI), revealed new details about the company’s strategy and eventual business model during a rare podcast appearance with Dwarakesh Patel.
Despite SSI raising nearly $3 billion since 2024 and still having no product, Sutskever insisted that monetisation will only follow after the core research is solved.
“Right now, we just focus on the research, and then the answer to this question will reveal itself,” he said. “I think there will be lots of possible answers.”
SSI began with the philosophy of “straight shot superintelligence”: prioritise building a safe superintelligent system first, and worry about productisation later. But Sutskever now says even that approach may require exposure to the public before reaching the finish line.
Patel asked why SSI would attempt to build superintelligence directly, while competitors release “weaker and weaker intelligence” models that gradually acclimate the public.
Sutskever agreed that releasing powerful AI incrementally may be necessary because society cannot meaningfully grasp its impact from essays and forecasts alone.
“It is nice to say, we’ll insulate ourselves from all this and just focus on the research and come out only when we are ready and not before,” he said.
But, he added: “I think on this point, even in the straight shot scenario, you would still do a gradual release of it… The gradualism would be an inherent component of any plan. It’s just a question of what is the first thing that you get out of the door.”
On the question of compute spending, Sutskever, who has been a vocal critic of the idea that simply “throwing more and more compute” will get the industry to superintelligence, also argued that SSI does not need the same scale of hardware investment as today’s AI giants.
He claimed that most of the enormous budgets at labs like OpenAI and Anthropic are tied up in inference, multimodal systems, staffing and product engineering — not in pure research.
“Then when you look at what’s actually left for research, the difference becomes a lot smaller,” he said. “The other thing is — that if you are doing something different —do you really need the absolute maximal scale to prove it? I don’t think it’s true at all.”
Patel challenged him further: if SSI is exploring “50 different ideas,” how can it know which ones will rival a breakthrough like the transformer without massive compute?
Sutskever responded: “I think that in our case, we have sufficient compute to convince ourselves and anyone else that what we’re doing is correct.”
Sutskever stressed that SSI’s edge is simply that it is pursuing a different paradigm, not trying to beat other labs on product cycles. “We are squarely [an] age of research company,” he said, and the goal is to first validate new ideas about generalisation before deciding how to deploy them.
Sutskever co-founded OpenAI in 2015 after leaving Google Brain, where he contributed to the breakthroughs that led to large-scale deep learning systems.
He exited OpenAI in May 2024 amid concerns that commercial pressure was overtaking the company’s original safety-first mission.
On June 19, 2024, he launched SSI with former Apple AI lead Daniel Gross and former OpenAI researcher Daniel Levy.
Operating between Palo Alto and Tel Aviv, SSI secured $1 billion by September 2024 and raised another $2 billion in April 2025, reportedly reaching a valuation of $30–32 billion despite having no product.
In July 2025, Gross left to join Meta’s AI division. Sutskever then took over as CEO, continuing SSI’s research-first push toward building safe superintelligence.
The post Ilya Sutskever Changes His Mind About How to Build Superintelligence appeared first on Analytics India Magazine.
