Ilya Sutskever's Safe Superintelligence (SSI) has raised roughly $8 billion and shipped exactly nothing, a gap between capital and product that Silicon Valley is watching closely. Sutskever helped build the technology that became ChatGPT, left OpenAI in 2024, and started SSI with a one-line mission statement. On July 27, 2026, Nvidia announced a $5 billion equity investment in SSI, confirmed in an official statement from Nvidia's newsroom, giving SSI priority access to Nvidia's next-generation Vera Rubin compute platform and expected to increase the startup's available compute by roughly tenfold over the next year. The Financial Times had already reported SSI closing an additional $2 billion earlier this year at a $32 billion valuation. Sutskever has said the company's first product would be safe superintelligence itself, and that it would not do anything else until then. In a 96-minute conversation with podcaster Dwarkesh Patel published in November 2025, he said "The data is finite" and "There is only one internet. Pre-training as we have known it is over," describing AI's history in three acts: a research era from 2012 to 2020, a scaling era from 2020 to 2025, and a new research era now beginning. In early August, investor Gavin Baker said in an interview with Patrick O'Shaughnessy that SSI planned to release its first