← United StatesThe Exclusive FeatureShare your journey
The people who built America’s companies · November 2026

Frontier AI research (safe superintelligence)

Safe Superintelligence

Sutskever argued a lab with one goal and no product cycle could chase safe superintelligence directly.

Ilya Sutskever
Ilya Sutskever. Photograph: Eladkarmel, CC BY-SA 4.0, via Wikimedia Commons

When Ilya Sutskever was sixteen, his family moved to Canada, and he lasted one month in high school. Then the University of Toronto admitted him as a third-year undergraduate. The detail sounds like a brag, but it also describes a boy who had changed countries twice.

He was born in 1986 in Gorky, in the Soviet Union, to a Russian-Jewish family. At five he emigrated to Israel and grew up in Jerusalem. A good student, he started taking classes at the Open University of Israel in eighth grade. As a teenager he reportedly wondered how learning works, and believed the mystery was not limited to humans. His father was an engineer who took him, at about five, to an expo, where the technology captivated him.

Within days he expressed regret and joined the employee letter demanding the board resign, the very board he sat on.

In Toronto he studied mathematics, graduating in 2005, took a master's in 2007 and finished a PhD in 2013 under Geoffrey Hinton. In 2012 he built AlexNet with Hinton and Alex Krizhevsky, the image-recognition network that helped start the deep-learning era. While still a student he visited DeepMind and turned down a job there, thinking the researchers had lost touch with reality because nobody was seriously discussing AGI. Google hired him, and at the end of 2015 he left to co-found OpenAI as chief scientist.

The next scene is the one that defines the public story. In November 2023 he voted, as a board member, to remove Sam Altman. At an all-hands meeting he called it the board doing its duty. Within days he expressed regret and joined the employee letter demanding the board resign, the very board he sat on. He reportedly had compiled a 70-page dossier on Altman before the vote. Altman was back within a week. In May 2024 Sutskever left.

On 19 June 2024 he announced Safe Superintelligence with Daniel Gross and Daniel Levy, promising that the first product would be safe superintelligence and nothing before it. In September 2024 it raised $1 billion at about $5 billion. A Greenoaks-led round in 2025 priced it near $30 billion, with no revenue. Meta tried to buy it and was turned away. In July 2025 Gross left for Meta, and Sutskever became chief executive. Nvidia announced a partnership on 27 July 2026, with a reported $5 billion investment the companies have not confirmed.

Sutskever has left two countries and one company already. The record suggests a man who has learned that the place that admits you is not always the place you belong, and who keeps starting over.

The Chronology year by year

  1. 2023
    • Sutskever votes to remove Altman, later reverses
  2. 2024
    • leaves OpenAI; 19 Jun 2024: SSI announced
    • $1B round
  3. 2025
    • Tel Aviv office
    • early 2025: rebuffs Meta takeover
    • Gross leaves, Sutskever CEO; 27 Jul 2026: Nvidia partnership

The Money Trail how it was funded

  1. Sep 2024 $1B (SV Angel, DST Global, Sequoia, a16z) at about $5B
  2. Mar-Apr 2025 round led by Greenoaks at $30B (Wikipedia) or $32B (other reports)
  3. Apr 2025 Google Cloud TPU deal
  4. 27 Jul 2026 Nvidia partnership with a press-reported $5B equity investment (terms undisclosed)

Principal backers: Greenoaks, Sequoia, Andreessen Horowitz, DST Global, SV Angel, Alphabet, Nvidia

The Close Calls what nearly broke it

  • 2025: Meta tried to acquire SSI and then hired Daniel Gross; staff poaching in 2025; no product or revenue yet

In Their Own Words from the record

Spotted something wrong? Send a correction.

Startup Talk · Preview edition · Figures are from public sources and may be out of date
About · Contact · Terms · Privacy · Badge · Share your journey