NVIDIA CEO Jensen Huang has directly rejected the growing “circular financing” narrative surrounding the company’s expanding investments across the artificial intelligence ecosystem. Speaking at the Goldman Sachs Communacopia + Technology Conference, Huang argued that NVIDIA’s capital deployments are small relative to the revenue they help generate, summarizing his view as: “I put in one, and a hundred comes back.” He stressed the figure was rhetorical, not a disclosed 100-times return, and insisted the arrangement is not circular because the company puts in a little money and a lot returns.
The debate centers on deals with major AI players, including OpenAI and CoreWeave. NVIDIA is both an investor in and partner to OpenAI, and it supports CoreWeave’s infrastructure. In January, Nvidia put $2 billion into CoreWeave Class A shares at $87.20 per share, alongside plans to help build more than 5 gigawatts of AI factories by 2030. Separately, OpenAI assembled a $110 billion investment round at a $730 billion pre-money valuation, including $30 billion from NVIDIA, $30 billion from SoftBank, and $50 billion from Amazon. NVIDIA has by now invested roughly $50 billion in AI labs and agreed to provide up to $105 billion in guarantees linked to OpenAI’s Ohio data-centre project, where it is the exclusive chip supplier. Its SEC filings also showed three direct customers accounted for 16%, 15%, and 13% of revenue in the first half of fiscal 2027.
Regulators and financial institutions are paying close attention. The IMF warned in April that AI-linked investments may face trouble in a downturn and flagged more circular financing across the AI value chain, though it called the current financial stability impact insignificant. The Bank for International Settlements pointed on September 10 to rising use of debt and private credit for AI capital outlays, warning that the investment boom could produce a much larger financial crisis if returns do not align. NVIDIA’s financing platform with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR aims to mobilize more than $500 billion in third-party capital, although that is a financing target, not an NVIDIA commitment.
Despite Huang’s defense, Nvidia’s stock still fell 2.37% to $218.36 on September 10, its third consecutive decline, though the Nasdaq’s 0.65% drop and Treasury yields near 5% indicated broader market pressure. Goldman Sachs maintained a Buy rating and $300 target on Nvidia, and Nvidia expects revenue growth of about 70% in its next fiscal year. Yet investor Dan Niles warned that hyperscalers may be double-ordering because they expect only part of their chip requests to arrive, complicating how demand should be interpreted.
The broader stakes are huge. S&P Global estimates the five largest hyperscalers may spend an additional $5.3 trillion on capital expenditures by 2030. Stanford University’s 2026 AI Index estimates global AI compute capacity reached 17.1 million H100-equivalents, with NVIDIA responsible for more than 60% of the total. Huang has also been narrowing expectations for months: in February he said of a possible OpenAI investment, “It was never a commitment,” and that NVIDIA would invest “one step at a time.” The distinction matters because if financing, orders and valuations run ahead of real revenues, the semiconductor, cloud, data centre and AI startup markets could face significant trouble.
