Nvidia Corp. (NASDAQ: NVDA)’s plan to mobilise more than $500 billion in third-party capital for artificial intelligence infrastructure has prompted fresh debate over how the next phase of the AI build-out will be financed, with one financial adviser arguing that investors should look beyond the headline size of the programme.
Nigel Green, chief executive of deVere Group, said the structure of the financing deserves closer examination because Nvidia has agreed to guarantee up to 25% of the residual value of its chips in individual financing transactions.
The arrangement involves memorandums of understanding with Apollo Global Management Inc. (NYSE: APO), BlackRock Inc. (NYSE: BLK), Blackstone Inc. (NYSE: BX), Brookfield Corp. (NYSE: BN), Goldman Sachs Group Inc. (NYSE: GS) and KKR & Co. Inc. (NYSE: KKR). The objective is to channel large amounts of private capital into AI infrastructure projects built around Nvidia hardware.
Green said the residual-value guarantee is the element investors should focus on.
“If demand for Nvidia’s chips is genuinely as strong and durable as the market currently believes, why does the company need to personally guarantee the resale value of its own hardware to get lenders comfortable?” he said.
The financing structure is intended to support the rapid expansion of AI data centres and computing infrastructure, which require significant upfront capital. Private credit investors and asset managers have increasingly become involved in funding these projects as demand for accelerated computing systems has grown.
Nvidia’s guarantee is designed to reduce lenders’ exposure to the future value of the chips if they are resold after the financing period. The company has introduced new GPU architectures roughly every two to three years, meaning lenders financing hardware over longer periods must consider how technological advances could affect the value of existing systems.
Green argued that this depreciation risk is central to understanding the financing.
“A lender being asked to finance chip purchases over a much longer horizon needs real confidence that the hardware retains value well beyond that replacement cycle,” he said.
He suggested that Nvidia’s willingness to guarantee part of the residual value could indicate that lenders would otherwise demand additional protection before extending credit at the required scale.
The comments reflect a broader debate over whether the current pace of AI infrastructure investment can continue without increasingly sophisticated financing structures. AI developers, cloud providers and data-centre operators are committing hundreds of billions of dollars to computing capacity, while equipment suppliers such as Nvidia are seeking to ensure that customers can finance those purchases.
Supporters of the approach argue that it can accelerate infrastructure deployment by reducing financing barriers and matching long-lived assets with institutional capital. They also note that Nvidia has a strong balance sheet and an established secondary market for its products, making a residual-value guarantee a potentially manageable commercial tool.
Green, however, compared the underlying mechanics to previous structured-finance arrangements in which one party provided assurances about the future value of the assets being financed. He stressed that the comparison was not intended to suggest that the same outcome would occur, but rather that investors should apply similar scrutiny to the incentives and risks involved.
“This does not mean history repeats exactly, but it means the underlying mechanics deserve the same level of scrutiny they would receive in any other sector,” he said.
He also pointed to a recent moderation in retail investor sentiment around Nvidia as an indication that some market participants are beginning to question the scale and structure of AI financing. He described the shift from bullish to more neutral sentiment as an early signal rather than a conclusion about the company’s prospects.
Nvidia remains the dominant supplier of AI accelerators used in many large-scale computing systems, and demand for its chips has driven substantial investment across the technology sector. The company’s financing initiative is intended to support continued expansion of that infrastructure by bringing private credit and institutional investors into the funding process.
Green said the programme should not be interpreted as evidence that AI infrastructure demand is weakening or that Nvidia faces immediate financial difficulties. Instead, he argued that the involvement of major private credit firms and the use of residual-value guarantees highlight how capital-intensive the next phase of AI deployment has become.
For investors, the question is less about whether AI infrastructure will continue to expand and more about the financial structures being used to sustain that expansion. Nvidia’s $500 billion financing platform brings together some of the world’s largest alternative asset managers and investment banks, underscoring the scale of capital required to build the computing infrastructure behind the AI boom.
Whether the arrangement ultimately becomes a model for financing future AI projects or simply a temporary solution to meet extraordinary demand will depend on how quickly the infrastructure is deployed, how long the hardware retains its economic value and whether the expected returns from AI computing justify the investment.
