Ilya Sutskever's Safe Superintelligence Inc. and NVIDIA Announce Long-Term Strategic Partnership

NVIDIA and Safe Superintelligence Inc. (SSI), the startup co-founded by Ilya Sutskever, announced a long-term strategic partnership on 27 July 2026. According to the statement, NVIDIA has also made an investment in SSI which, combined with access to its upcoming Vera Rubin compute platform, will allow SSI to increase its compute capacity by an order of magnitude (that is, roughly tenfold). The statement does not specify the exact size of the investment.

Both companies note that they will also collaborate on the technical development of NVIDIA's current and future compute platforms, drawing on what the statement describes as SSI's "unique insights" into the future of AI. In other words, the relationship is not framed as a simple hardware sale, but as a two-way collaboration: NVIDIA contributes infrastructure and capital, and SSI contributes research information that could shape the design of future chips and systems.

The text indicates that SSI has spent two years working quietly on a new line of research aimed at achieving powerful and "robustly aligned" artificial intelligence (that is, safety and control over the system's behavior). NVIDIA says it was granted unusual access to that research, normally closely held, before deciding to enter the partnership to accelerate SSI's next stage of growth.

In remarks quoted in the statement itself, Jensen Huang, founder and CEO of NVIDIA, highlighted Sutskever's track record as a pioneer of fundamental advances in modern AI, specifically mentioning AlexNet as a starting point, and expressed interest in the discoveries SSI may achieve with the support of the Vera Rubin platform. For his part, Ilya Sutskever, co-founder and CEO of SSI, said the company has a line of research it considers "worth scaling," and that access to a large NVIDIA compute system will make that possible; he added confidence that the bet on Vera Rubin will take SSI "to the next level."

On SSI, the statement recalls that it was founded in 2024 and is led by Ilya Sutskever and Daniel Levy, and describes itself as the first straight-shot SSI lab, with a single goal and a single product: a safe superintelligence. It details that Sutskever's earlier research laid part of the foundations of modern AI, with contributions to AlexNet, AlphaGo, Sequence-to-Sequence learning and the GPT models, as well as driving the research that led to reasoning models such as OpenAI o1. SSI's investors include Andreessen Horowitz, DST Global, Greenoaks and Sequoia Capital.

The announcement comes amid a series of recent NVIDIA moves to build AI infrastructure alliances with various players — the statement itself links to other news from those same days about "AI factory" expansions with SK Group, NAVER/Brookfield in Korea and a joint lab with KAIST — suggesting an NVIDIA strategy of securing deep relationships (investment plus compute supply) with research labs deemed strategically relevant, not just with large-scale cloud customers.

It is worth noting the limits of what the statement allows one to assert with certainty: the amount of NVIDIA's investment in SSI is not disclosed, no technical specifications for Vera Rubin or availability timelines are given, and there is no explanation of exactly what SSI's robust alignment research consists of beyond the generic description offered. The entire text comes from a joint press release, so its tone is promotional and should be read with the usual caution toward such sources, including the standard forward-looking statements clause the release includes.