A Partnership Built on Faith, Not a Product
Safe Superintelligence is the most extreme case study in how the AI industry allocates capital in 2026. On 27 July 2026, Nvidia and SSI announced a long-term strategic partnership that, according to reporting on the deal, involves roughly $5 billion of investment — a mix of cash and committed compute — and priority access to Nvidia’s next-generation Vera Rubin platform, with the arrangement expected to expand SSI’s available computing capacity roughly tenfold over the following twelve months.
What makes the deal remarkable is what SSI has not done. The company, founded in June 2024 by former OpenAI chief scientist Ilya Sutskever alongside Daniel Gross and Daniel Levy, has never shipped a product, published a research paper, or disclosed any revenue. It employs roughly 50 people. Yet it carries a $32 billion valuation, and one of the most valuable companies on earth just wrote a nine-figure-plus cheque to deepen its stake. Nvidia is not buying a revenue stream; it is buying a claim on a future it wants to help build.
The Numbers Behind the Valuation
SSI’s funding history reads like a chart of accelerating conviction. The lab raised $1 billion in September 2024 at roughly a $5 billion valuation, backed by Andreessen Horowitz, Sequoia, DST Global and others. Seven months later, in April 2025, it raised about $2 billion at a $32 billion valuation in a round led by Greenoaks Capital — a more than sixfold jump in headline value before any product existed. That $32 billion figure has held as the reference price ever since, and the July 2026 Nvidia partnership was structured as a strategic infusion rather than a fresh repricing.
Tallying disclosed capital is itself an exercise in ambiguity. Depending on how the Nvidia commitment and compute-in-kind are counted, reporting places SSI’s total funds somewhere between roughly $6 billion and $8 billion. Against roughly 50 employees, that implies a capital intensity on the order of $160 million per head — a figure that captures, better than any valuation multiple, how far the AI market has moved from traditional revenue-based pricing.
Advertisement
The “Straight Shot” Doctrine
The strategy Sutskever articulated is deliberately austere. SSI has described itself as the world’s first “straight-shot” superintelligence lab, with “one goal and one product” — a safe superintelligence — and has said it will not release interim products, APIs, or chatbots until that goal is reached, insulating research from “management overhead” and “product cycles.” Sutskever’s stated position is that the company’s first product will be safe superintelligence itself, and that it “will not do anything else up until then.”
That doctrine is a direct repudiation of the prevailing playbook. OpenAI, Anthropic and Google DeepMind all ship steadily, monetize aggressively, and iterate in public. SSI is betting that a single, undistracted push toward a transformative system will outrun the incrementalists — and it has persuaded some of the most sophisticated investors in technology to underwrite that bet with billions, entirely on the strength of a founding team’s track record. Whether a first model actually surfaces has become the industry’s most-watched open question: investor Gavin Baker noted in early August 2026 that SSI had signaled a model “in August,” though as of early August 2026 the company issued no formal confirmation.
The Circular-Financing Question
The Nvidia deal cannot be read in isolation, because it is one node in a dense web of interlocking AI investments that drew intensifying scrutiny through August 2026. Nvidia’s aggregate AI-infrastructure commitments reportedly exceed $750 billion in deals in motion, spanning stakes and compute arrangements with OpenAI, Anthropic, CoreWeave and others. Critics describe a pattern of “circular financing”: Nvidia invests in AI companies, which then use that capital to buy Nvidia chips, which inflates Nvidia’s revenue, which supports the valuation that funds the next round of investments.
The SSI partnership fits the template neatly — Nvidia supplies both capital and the compute that capital is earmarked to consume. Defenders argue this is ordinary strategic investment that seeds demand for a genuinely scarce product, not accounting sleight-of-hand. Skeptics counter that the interlocking relationships create contagion risk: if demand forecasts fall short, impairment at one node could ripple across the whole ecosystem, echoing the vendor-financing dynamics that broke telecom during the late-1990s fiber build-out. Both readings are defensible, and the honest position in August 2026 is that the model has not yet been tested by a downturn.
Why This Is a Strategic Signal, Not Just a Big Cheque
Strip away the personalities and the SSI deal is a statement about how the frontier of AI is now financed. The market is pricing three things that traditional valuation ignores: the scarcity of elite research talent, the strategic value of guaranteed compute access, and a belief that whoever reaches transformative AI first captures a disproportionate share of the value. On all three, SSI scores highly despite having nothing to show a customer — because its founder co-created the modern deep-learning era and its compute is now underwritten by the company that makes the chips everyone else must queue for.
That is also the risk. A $32 billion valuation with zero external validation means the entire thesis rests on unaudited conviction. Neither the Nvidia partnership nor the multi-billion-dollar cash pile solves the underwriting problem of having no outside observer able to validate the work. Whether $32 billion proves cheap or absurd will depend on whether a release eventually answers the question the market has so far been content to leave open. For anyone tracking AI strategy, SSI is the purest available signal of where sentiment sits — and the clearest test of whether conviction-based pricing survives contact with results.
Frequently Asked Questions
What is Safe Superintelligence and who founded it?
Safe Superintelligence (SSI) is an AI lab founded in June 2024 by Ilya Sutskever, the former chief scientist and co-founder of OpenAI, alongside Daniel Gross and Daniel Levy. Its stated goal is to build a single product — a safe superintelligence — through a “straight-shot” research strategy, releasing no interim commercial products until that goal is reached.
How much did Nvidia invest and what did SSI give up?
Nvidia announced a long-term strategic partnership on 27 July 2026 that reporting places at roughly $5 billion in cash and committed Vera Rubin compute, expected to increase SSI’s computing capacity about tenfold over twelve months. The deal deepened Nvidia’s existing stake; SSI’s valuation remained at $32 billion, the reference price set in its April 2025 round.
Why is a company with no product worth $32 billion?
The valuation reflects the scarcity of elite AI research talent, the strategic value of guaranteed compute, and a belief that reaching transformative AI first captures outsized value. Critics counter that $32 billion with zero revenue, no product and no published research rests entirely on unaudited conviction, and that Nvidia’s web of interlocking “circular financing” deals adds contagion risk if AI demand forecasts fall short.
Sources & Further Reading
- Ilya Sutskever’s Safe Superintelligence Inc. and NVIDIA Announce Long-Term Strategic Partnership — NVIDIA Newsroom
- Ilya Sutskever’s Safe Superintelligence partners with Nvidia to scale its AI research — TechCrunch
- This AI startup has roughly 50 employees, no product and has never published a word of research. It’s worth $32 billion — Yahoo Finance / Fortune
- Ilya Sutskever’s Safe Superintelligence raises $2B at $32B valuation — with no product yet — CTech
- OpenAI co-founder Ilya Sutskever’s Safe Superintelligence reportedly valued at $32B — TechCrunch
- Is Nvidia Engaging in Circular Financing? Some Think So — The Motley Fool
- Nvidia is about to spend $750 billion on AI. Critics are calling it a bubble — NPR
- SSI Valuation 2026: $32B Mark, $8B Raised, Nothing Shipped — ValueAdd VC
