In an attempt to facilitate the buildout of AI infrastructure over time, Nvidia and six financial partners said they are creating a fund of more than $500 billion of third-party capital.
The impact of such a cash infusion on enterprise AI is uncertain, but analysts fear that it could both further increase enterprise AI infrastructure costs and exacerbate the shortage of AI chips for data centers.
The announcement from Nvidia and financial partners Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR said that their memorandums of understanding describe a fund “to establish the first compute financing platforms of their kind at global scale to enable the AI infrastructure buildout across Nvidia’s ecosystem, including leading frontier AI labs, enterprises and AI clouds.”
The group added that the fund would “create dedicated pools of capital at significant scale at attractive rates for Nvidia customers.”
Although the statement said the goal was to help AI infrastructure “across Nvidia’s ecosystem, including leading frontier AI labs, enterprises and AI clouds,” analysts and consultants agreed that it is highly unlikely any of these funds would be dispensed directly to enterprises, but would instead impact the overall AI supply chain.
Even the precise amount of money earmarked for the fund was unclear, with the statement merely saying that the amount would be more than $500 billion.
Nvidia did not respond to requests for clarification about details of the proposed fund, but one financial partner did comment on the amount.
“We can clarify that this is a number that’s been totaled up by Nvidia,” said Simon Maine, managing director for communications at Brookfield Asset Management, in an email. “The finance partners are not collaborating together on this, but rather it is a series of individual partnerships. We therefore cannot comment on how the total figure has been arrived at.”
The top concern for CIOs around such a fund is the question of whether it would impact chip pricing along with that of components and devices using those chips, and if it would impact chip availability. Almost all of the analysts and consultants willing to speculate on that agreed that it would likely increase prices and worsen chip shortages. However, one optimistic interpretation of the fund was that it could help reduce chip shortages.
“It remains to be seen what kind of downstream impact this initiative will have on enterprise spend,” said Ashish Nadkarni, a group VP for IDC, but “there is an assumption here that these investments will go toward building fab capacity, and that the fabs will produce chips to address a chip shortfall.”
Other analysts disagreed, and argued that the fund would likely make the chip shortage worse, at least initially.
“The current chip demand is taking all of the capacity and there is only so much chip fabrication capacity available,” said Mark Tauschek, a distinguished analyst at Info-Tech Research Group. “It will also take years to build new chip fabs. [The proposed Nvidia fund] will probably exacerbate the shortage.”
In fact, for enterprises, Tauschek projected a 15%-20% cost hike.
Sanchit Vir Gogia, chief analyst at Greyhound Research, agreed with Tauschek on both counts.
He estimated that AI chip prices would be roughly the same for another year, and increasingly only at a good price if customers sign a long-term commitment. “The discount for committing is shrinking, not growing,” he said.
Thus, he pointed out, “more financing therefore means more new capacity is spoken for before it exists, and the open market gets whatever is left. The queue is no longer sorted by who can pay. It is sorted by who will commit.”
But he added that these numbers will likely improve eventually. “It does add real capacity in the end,” he noted. “Enterprises planning for 2028 will benefit. Those reacting to 2026 will not.”
Mike Wilkes, enterprise CISO at Aikido Security, said that one of the key impacts of the fund will be the way in which enterprises should view AI financing. That change, he argued, is both good and bad.
The improved availability of funds could “finance the AI buildout at much greater scale. That could accelerate enterprise access to compute, but it could also connect AI infrastructure much more tightly to the financial system,” Wilkes said. “This financing is likely to lower the cost of getting access to AI infrastructure in the near term, but not necessarily lower the price enterprises ultimately pay for AI.”
He suggested that potential enterprise impact will vary over time.
For large enterprises, the infusion of funds would allow their infrastructure vendors to build large datacenters without requiring financial help from the enterprise, he said, and this added capacity should eventually put downward pressure on the unit cost of compute. “But,” he noted, “in the next few years, I would expect vendors to use cheaper financing primarily to build faster and lock customers into longer-term capacity contracts, rather than simply pass all of those savings through. In other words, enterprises may get more AI for the same dollar before they get the same AI for fewer dollars. So I think the biggest effect on enterprise readers is that this could remove one bottleneck while creating another.”
He expects that capital may cease to be the limiting factor in building AI infrastructure, giving CIOs considerably more capacity and more financing options available to them. “But,” he said, “the resulting competition may increasingly focus on who can persuade enterprises to make the longest and largest commitments to future AI consumption. If hundreds of billions of dollars of infrastructure are financed based on assumptions about future utilization, somebody ultimately has to pay when those assumptions prove wrong.”
Justin Greis, CEO of consulting firm Acceligence, agreed that the short-term impact of this agreement may not be good for enterprise IT.
“My view is that this financing will accelerate the creation of AI infrastructure, but it will not provide meaningful near-term price relief for most enterprises. In fact, I think the next 12-18 months could remain a period of elevated costs and constrained availability as the market absorbs this new wave of investment,” Greis said. The reason, he noted, is that the limiting factor today is not capital, it is the physical capacity to manufacture components.
“Adding hundreds of billions of dollars of available financing will create more buyers with the ability to compete for those resources,” he said. “My expectation is that the largest AI infrastructure providers and hyperscalers will continue to secure a significant share of available capacity because they have the scale, existing relationships, and ability to commit to long-term purchases.”
But eventually he believes that the enterprise picture should brighten.
“Where I do expect enterprises to see benefits is further out. As this capital turns into actual infrastructure capacity, the market should become more competitive and enterprises should have more options in how they access compute,” Greis said.
“From a CIO perspective, I would not respond to this announcement by trying to secure more hardware. That is likely to become an expensive race that most enterprises cannot win.”