In June 2024, Ilya Sutskever, Daniel Gross, and Daniel Levy founded Safe Superintelligence Inc. (SSI), which quickly became noticeable in AI. SSI, based in Palo Alto, California, and Tel Aviv, Israel, has one clear goal: to build a superintelligence that is smarter than humans and still focuses on safety. This paper describes the history behind SSI, its mission, where it gets its funding, and what its role in AI safety could be.
Table of Contents
Founding Vision: A Safety-First Stand in a High-Stakes Industry
A group of renowned experts in AI started Safe Superintelligence Inc. With his prior achievements like AlexNet in 2012 and ChatGPT, Ilya Sutskever is recognised as a leading scientist for moulding key developments in AI. Sutskever left OpenAI in May 2024 after voting to terminate the CEO, Sam Altman, because he felt that the company’s focus was moving from prioritising safety to making profits. Daniel Gross, Apple’s former head of AI and co-founder of Cue and Daniel Levy, who was at OpenAI and invests in AI, are also joining him. Because of its team’s experience, SSI is well placed to compete for the development of advanced AI technologies.
A Mission Beyond Profit
SSI’s mission is to create a safe superintelligence. Unlike other AI firms focused on profit, SSI prioritises AI safety. The company combines safety and capability by promoting swift growth in AI while keeping safety measures ahead. This approach allows SSI to focus on long-term research and development without the pressure of short-term profits.
Sutskever compares SSI’s safety perspective to the nuclear industry’s approach to radiation, rather than the typical online content moderation model. The primary goal is to ensure superintelligent systems do not endanger humanity, whether accidentally or due to differing values. By thoroughly testing its software, publicising its processes, and valuing expert opinions, SSI earns user trust more than its competitors and maintains honesty in its results.
Sky-High Valuations: Betting Big on a Safer AI Future
Although SSI is still a young company with only about 20 staff, investors are very interested. In the same month, September 2022, the company attracted $1 billion at a $5 billion valuation from leading venture capital firms such as Andreessen Horowitz, Sequoia Capital, DST Global, SV Angel and NFDG, an investment group formed by Daniel Gross and Nat Friedman. In March 2025, SSI’s value reached $30 billion after Greenoaks Capital led a funding round, and it grew to $32 billion within the next month after the company raised an extra $2 billion. Some financial backers behind SSI, including Alphabet and Nvidia, rely on Sutskever’s achievements and clear mission, even though SSI lacks income and a physical product. Most funds are put into getting powerful equipment and inviting world-class engineers and researchers.
Innovative Development Approach
SSI approaches things differently by avoiding the industry trend of releasing generative AI products quickly. OpenAI and Anthropic offer consumer models now, but SSI plans to spend a long time researching and developing before launching its superintelligence. According to this strategy, AI systems should be safe, and alignment is recognised as a main concern. Sutskever sees a link between AI and natural processes, suggesting that imitating how humans instantly make decisions can boost AI.
At the 2024 NeurIPS event, Sutskever emphasised the role of AI agents that can perform higher-level reasoning instead of big language models. As per this vision, SSI looks to change how models are scaled, focusing on structure so new solutions can develop rather than putting effort into only larger models. The organisation hopes to lead the way in AI safety and capability by having a select group of leading researchers from Palo Alto and Tel Aviv.
Competing in a Crowded Field
It operates alongside many companies in the AI industry. Experts believe that ensuring superintelligence is safe presents significant challenges, especially since AI struggles with common-sense questions. OpenAI earns substantial resources, while Anthropic focuses on building secure AI for companies. With only one product in development compared to established portfolios, SSI has a distinct approach that raises doubts about its timelines and novelty. Analyst Chirag Mehta noted that predicting SSI’s future is difficult; its success will depend on attracting exceptional talent and forming important alliances.
SSI’s existence highlights ongoing challenges within AI regarding the balance between safety and profitability. Sutskever’s departure from OpenAI and other safety-focused researchers like Jan Leike’s move to Anthropic illustrate the growing divide between those wanting rapid product deployment and those prioritising safety. SSI’s mission critiques OpenAI’s recent direction, claiming it is reverting to its initial values of safety-focused AI research.
Challenging the Industry’s Priorities
If SSI’s goal of safe superintelligence is met, it could greatly change how artificial intelligence is developed. The company tries to drive the industry to accept stricter safety measures by prioritising safety and alignment. Because it has a large budget and an influential group of founders, and as it recruits staff in Palo Alto and Tel Aviv, investors are closely watching this company. Yet, not having a flat pathway for the project and the difficulty in achieving the goal of superintelligence mean there are risks. Since SSI is fighting against OpenAI, Anthropic and Google DeepMind, its success depends on showing significant progress in AI safety and technology.
A Risky Path, but One Worth Watching
Safe Superintelligence Inc. emphasises the importance of ensuring safety and alignment before developing any superintelligent system. Because SSI has a clear purpose, effective operations, and ample resources, it is ready to make valuable improvements to AI safety. As it advances, those involved in the technology and ethics fields will monitor the company’s work.
