Safe Superintelligence

Safe Superintelligence Inc. (SSI) is an artificial-intelligence research company founded in June 2024 by Ilya Sutskever, the former chief scientist of OpenAI, together with the investor and former Apple AI executive Daniel Gross and the former OpenAI researcher Daniel Levy. The company was announced with a one-sentence mission, that it had "one goal and one product: a safe superintelligence," and was structured deliberately to ship no commercial products until that goal is reached, insulating its research from the revenue pressures that shaped OpenAI and Anthropic. It operates from offices in Palo Alto and Tel Aviv.¹²

Founding and the No-Product Model

Sutskever left OpenAI on May 14, 2024, months after the November 2023 board crisis in which he had voted to remove and then reinstate chief executive Sam Altman, and announced SSI roughly a month later. His cofounder Daniel Gross had founded the personal-assistant startup Cue, sold it to Apple in 2013, run AI efforts there, and become a partner at Y Combinator before partnering with the former GitHub chief executive Nat Friedman to run the venture fund NFDG; Daniel Levy had led the optimization team under Sutskever at OpenAI. In its launch statement the founders framed the single-goal structure as a feature, writing that the company's "singular focus means no distraction by management overhead or product cycles, and our business model means safety, security, and progress are all insulated from short-term commercial pressures."¹³

The company has published no products, papers, or model releases of note, and its website functions as little more than the founding statement. The deliberate absence of a product made SSI an unusual fundraising case: investors backed a research program with no revenue, no public benchmarks, and an explicitly open-ended timeline, on the strength of Sutskever's record and the safety-first thesis.²⁴

Funding Rounds and Investors

In September 2024 SSI raised 1 billion dollars at a 5-billion-dollar valuation, in a round whose backers included Andreessen Horowitz, Sequoia Capital, DST Global, SV Angel, and NFDG, the fund run by Gross and Friedman. The three-month-old company had no product, yet the round placed it among the most richly valued AI startups of its cohort. Reporting in early 2025 indicated a far larger follow-on, and in April 2025 the company was reported to have raised an additional 2 billion dollars at a 32-billion-dollar valuation, led by Greenoaks, which put in a reported 500 million dollars, with participation from Andreessen Horowitz, Lightspeed Venture Partners, and DST Global.⁴⁵

Among the later backers were Alphabet and Nvidia, both strategic investors with stakes in AI infrastructure, and Google Cloud became a major compute provider to the lab, supplying it with Tensor Processing Units.⁴⁶

The Meta Approach and Sutskever as CEO

In June 2025 Meta Platforms, building its superintelligence division under Mark Zuckerberg, attempted to acquire SSI outright. Sutskever rebuffed the approach. Meta instead hired Gross, who had been SSI's chief executive, along with his NFDG partner Nat Friedman, in a transaction in which Meta took a stake in the NFDG fund; Gross departed SSI on June 29, 2025 to work on AI products at Meta. On July 3, 2025 Sutskever announced that he had become chief executive of SSI, with Daniel Levy as president and the technical staff continuing to report to him.⁷⁸

Sutskever ruled out any sale of the company, writing that "we are flattered by their attention but are focused on seeing our work through. We have the compute, we have the team, and we know what to do. Together we will keep building safe superintelligence." Gross, for his part, said as he left that he expected "miracles" from the lab he had cofounded. The episode left Sutskever as both the public face and the operating head of a lab that had been built around his name and continued to hold its 32-billion-dollar valuation without a product.⁷⁸