Nvidia invests in Sutskever's SSI to increase its computing power tenfold.

      Nvidia has acquired a stake in a company that currently has no product, no revenue, and no immediate plans to sell anything. Safe Superintelligence (SSI), established by Ilya Sutskever after his departure from OpenAI, announced on Monday that Nvidia had made a significant investment as part of a long-term collaboration. The exact amount was not revealed by either party.

      What the funds will achieve

      SSI explicitly stated the purpose of this investment. “NVIDIA is making a substantial investment in SSI that will let us 10x our compute in the next 12 months,” the company shared on X. Then came the crucial statement: “We reached the point where our research is worth scaling.” Sutskever added, “Time to scale that SSI.” Co-founder Daniel Levy contributed a more colorful take: “Deep learning happens when a small, cracked team operates a big computer. The computer just got bigger.” The agreement also grants SSI access to Nvidia’s Vera Rubin platform, which commenced full production this month, a significant detail since OpenAI is implementing the same generation on a large scale this quarter, and supply remains a critical limitation in the industry. A lab without a product is now in line with the organization managing the world's largest deployment.

      Why this statement is significant

      For two years, SSI has operated as one of the most discreet companies in AI. It has raised around $3 billion from Andreessen Horowitz, Sequoia Capital, and DST Global, achieved a valuation of $32 billion, and produced nothing. This is intentional, as the company has declared it does not plan to offer AI products in the near future. It has a singular focus and lacks a roadmap aimed at customers. Sutskever left OpenAI in May 2024, having played a significant role in its foundational research, and he revealed a $7 billion stake during the Musk litigation; thus, he is not seeking funding out of necessity. His public remarks have been cautious, and in November he clarified a summary of his own podcast, noting that scaling would yield ongoing improvements without plateauing, but indicated that “something important will continue to be missing.” Viewed in this context, the statement “our research is worth scaling” suggests that some of the previously missing elements may now exist. However, verification remains impossible for anyone outside of SSI, complicating matters for the company.

      Nvidia’s due diligence

      One noteworthy aspect reported by the Wall Street Journal is that Nvidia invested after gaining rare insight into SSI’s research state. A company that typically reveals nothing shared some information with its chip supplier. What Nvidia discovered is not public knowledge and may never be revealed. The fact that Nvidia scrutinized SSI before investing is significant and distinguishes this deal from a simple financing arrangement.

      A recognizable trend

      This is not Nvidia's first investment in a research lab; it previously supported David Silver’s Ineffable Intelligence with $5.1 billion, alongside Sequoia, another lab centered on a notable researcher rather than a product. Its competitor, AMD, is similarly investing up to $5 billion in Anthropic and deploying two gigawatts of its hardware to run Claude. The Journal also suggests that the SSI deal serves as a defensive measure to prevent a prominent customer from falling into a rival's hands, indicating that chip manufacturers are increasingly investing in labs rather than just supplying them.

      The visible cycle

      Bloomberg clearly articulates the concern that such deals have faced criticism for their circular nature, intertwining suppliers and startups in mutual dependence. SSI exemplifies this scenario, having no customers and generating no revenue. The only way it will utilize Nvidia’s investment is to purchase and operate Nvidia hardware. Additionally, Nvidia is in discussions to secure up to $250 billion in financing for OpenAI data centers, a report released a day prior to this news.

      Market reactions

      Investors did not respond positively; Nvidia’s shares dropped as much as 2.3% to $202.13 on Monday, according to Bloomberg. This aligns with recent trends where announcements of significant AI investments, which once boosted share prices, are now negatively impacting them. The pressing issue shaping the markets is not merely how much a company is investing in AI but how much of that investment translates into revenue from external sources.

      Two distinct wagers

      Investors in SSI are wagering that a small team with robust computing resources will create something unique. In contrast, Nvidia’s bet is more targeted and considerably safer; whatever SSI discovers will be based on Nvidia chips. Ultimately, only one of these bets can yield a successful outcome.