Nvidia chief executive Jensen Huang went on television Wednesday to argue that the chipmaker’s expanding role as financier to the AI boom is not a round-trip accounting trick but a structural necessity, a distinction that matters because the company has now committed or arranged more than half a trillion dollars in capital to the very customers buying its chips.
The circular financing charge
Critics have labeled the arrangements circular financing: Nvidia provides funding to data center developers and model builders, who then spend a portion of that capital on Nvidia hardware, inflating the chipmaker’s revenue while keeping the risk on its own balance sheet. The comparison to dot-com era vendor financing is explicit in the market chatter, and the scale is without recent precedent, $105 billion backing a single Ohio campus leased to OpenAI, plus a new syndicate with Wall Street banks targeting up to $500 billion for data center builds.
The capital intensity argument
Huang’s rebuttal on CNBC rested on a single premise: frontier AI companies require tens of billions of dollars before they generate a dollar of profit, and none of them are investment grade. “They don’t have the track record, the capital track record, the financial track record, to be able to capture or secure capital at a low cost,” he said. Nvidia’s balance sheet, swollen by the same demand it is now underwriting, becomes the bridge. The source does not disclose whether Nvidia’s stakes in OpenAI, Anthropic, or the neocloud providers are equity, convertible debt, or revenue-linked instruments, nor does it state any break fees or collateral terms on the Ohio or syndicated facilities.
The redeployment thesis
The CEO argued that the underlying compute infrastructure is fungible, if one tenant fails, the servers can be re-leased to another workload, limiting single-name exposure. “The money we’ve invested is going to generate tremendous returns,” Huang said. “I think the risk is low.” That claim sits alongside a quarter that delivered $96.2 billion in revenue, more than double the year-ago period, with data center sales up 117 percent to $89 billion. The company guided for roughly 70 percent revenue growth in the current quarter.
What the numbers actually show
The revenue acceleration suggests end-user demand remains real, but the financing layer means Nvidia is increasingly both the supplier and the bank. Whether that creates a self-reinforcing loop or a genuine capital formation function depends on whether the AI labs ever reach cash-flow breakeven without further vendor support. The source offers no timeline for that transition.