Artificial intelligence infrastructure has taken the decisive leap from being an investment niche to becoming an investment class in its own right.
Nvidia successfully raising a private capital pool of $500 billion from some of the world's biggest financiers proves the sector’s pulling power, which analysts say is only bound to grow stronger.
Jensen Huang, Nvidia's chief executive on Tuesday said he only approached six companies for the half-trillion-dollar fund, and none turned him down.
“AI infrastructure is becoming financeable through mainstream institutional channels rather than relying mainly on Nvidia, hyperscalers or sovereign investors,” Safwan Mirza, equity analyst at Mirabaud Wealth Management, said.
“Investors increasingly view AI infrastructure as an opportunity capable of generating long-term, recurring cash flows, not simply as a strategic technology project.”
The announced $500 billion is a potential investment capacity, not money that has already been committed or deployed.
However, Nvidia’s move has turned AI computing infrastructure into a “financeable asset class” just like car or aircraft leasing, or solar and wind farm projects, “with attractive financing for capital constrained data centre customers”, Cengiz Sen, equity research analyst at Swiss private bank Julius Baer, said.
Who’s lined up
Those who lined up to answer Mr Haung’s call are some of the biggest asset managers in the world. BlackRock, Apollo Global Management, Blackstone, and Brookfield Asset Management, collectively have around $19 trillion in aggregate assets under management as of the end of the second quarter.
The coalition also includes Goldman Sachs Group, one of the biggest investment banks in the world, as well as KKR, among the top global alternative investors.
The six of them will “create dedicated pools of capital at significant scale at attractive rates for Nvidia customers”, according to Nvidia.
The involvement of the big-name financiers will spur the race to join more such deals or be part of the existing ones. Bank of America analysts see more deals, either in debt or equity capital financing in the AI infrastructure space.
The global AI market’s potential annual revenue opportunity is estimated to reach around $1.8 trillion by 2030, while Goldman Sachs Research estimates annual AI-related investment of approximately $1 trillion globally in 2026.
“This indicates a potentially very large financing opportunity,” Mr Mirza said.
The focus
Mr Huang and top executives of the financing firms have said the focus of the debt financing is to provide access to compute for Nvidia’s largest customers.
The aim of bringing together the long-term capital providers is to “independently underwrite AI infrastructure” as those financing platforms will help customers access “scarce compute at scale and build the AI factories that will power every industry and country in the age of AI”, according to Mr Huang.
BofA Securities analysts said for $500 billion of capital to treat “compute as an investable asset class”, residual value must hold, and that is exactly what Nvidia supplies as it “guarantees asset quality”.
“Similar to a toll road, AI infrastructure is [now] viewed as a recurring cash flow facility, with customers paying for every AI workload. Investors can underwrite those future cash flows just as they finance toll roads, airports, or renewable energy projects,” Mr Sen from Julius Baer explained.
Flurry of deals
Nvidia's deal comes as sovereign investors and Wall Street institutions have been mobilising hundreds of billions of dollars in debt and equity deals for AI and related infrastructure development.
US firms such as Alphabet, Google, Microsoft, Meta and Oracle have tapped credit and equity markets for more than $450 billion since early 2025 and more equity raises are expected this year and next.
On the credit side, the largest transactions include Meta's $30 billion bond deal in October 2025, Amazon's $37 billion bond sale in March this year and a further $25 billion one in July, Alphabet's $27 billion multicurrency offering in February this year, and Oracle's $25 billion bond issuance during the same month. On the equity side, Alphabet executed the largest deal of $85 billion in June.
Separately, Nvidia has already been in talks to backstop as much as $250 billion to help OpenAI lease computing power from its $500 billion, 10-gigawatt data centre, easily the company’s biggest financing deals with a customer, according to media reports.
Nvidia was also in discussions to finance $350 billion of OpenAI’s purchases of its chips for the project, reports added.
In a post on X this week, Mr Huang said the company may provide some financing support of “up to 25 per cent of an opportunity” with the pack of six. “Our role is to help unlock a very large pool of independent capital while maintaining disciplined risk exposure,” he wrote.
Analysts say, so far, sources of capital for US tech giants are mostly banks, institutional investors and private credit.
“We have not seen many examples of private infrastructure funds providing capital to the hyperscalers so far, though they are involved in smaller developments and in the provision of services tangential to data centres – eg, providing power,” Enrico Chinello, analyst at Julius Baer’s Next Generation Research, said.
Sovereign funds
Sovereign funds and government related entities have also played their part, pouring billions through bilateral investment deals and alliances.
In 2024, Global Infrastructure Partners (GIP), a part of BlackRock, Microsoft, and Abu Dhabi’s MGX formed the AI Infrastructure Partnership, a consortium for AI infrastructure investments across the globe.
It aims to deploy $30 billion in equity, potentially reaching $100 billion including debt, to fund the global expansion of digital capacity.
In May last year, UAE President Sheikh Mohamed and US President Donald Trump witnessed the unveiling of plans for a new 5GW UAE-US AI Campus. The almost 26-square-kilometre campus is being built by UAE AI company G42 in Abu Dhabi and will be operated in partnership with several US firms.
Earlier this year, UAE-based sovereign AI infrastructure and platform company Aleria also announced it is building one of the world’s most powerful sovereign AI systems after securing a major deal with Nvidia for thousands of next-generation chips.
In July, G42 also said it will join the Nvidia-led Open Secure AI Alliance, teaming up with other companies in endorsing open-weight and open-source AI models that can be freely downloaded, shared and adapted.
Mr Mirza of Mirabaud said sovereign wealth funds will likely remain an important part of the AI investment scene in the future as they can provide large amounts of long-term capital. But they increasingly will operate alongside private institutional investors.
“This combination could be complementary: sovereign investors may provide long-term equity, while private funds, infrastructure investors and lenders provide additional capital and project-level discipline,” he said. “The involvement of independent investors may also help address concerns about circular financing.”
Risk factors
In terms of Nvidia’s pact, not enough is known beyond the headline figure so far. No details have been given for the timeline for the fund, or how potential investment financing deals will be structured. The scope of deals is also undisclosed.
Nvidia has already signed hundreds of billions of dollars’ worth of deals with companies across the AI ecosystem, which is stoking concerns from some investors that the company is trying to sell more of its chips through the circular nature of such agreements.
“Nvidia keeps trying to help its customers find money to buy its chips – this is worrying, because it inflates demand for Nvidia’s products,” Vey-Sern Ling, senior equity adviser at private bank UBP, said.
Investors have already seen debt and equity financing to fund AI investments from companies like Alphabet, Amazon and Oracle eat into their free cash flow.
“The market does not like this because balance sheets weaken and ROI [return on investment] of AI spending remains uncertain,” Mr Ling said.
“What is clear is that AI infrastructure, semiconductors, or ‘picks and shovels’, will continue to be the biggest beneficiaries in such a buildout.”
AI infrastructure becoming a target amid rising geopolitical uncertainties is another risk investors have to contend with, according to analysts.
Mr Mirza said the more AI infrastructure is treated as strategic infrastructure, the more important political and operational risks will be in underwriting such projects.
“As geopolitical tensions rise, we could see an increasing number of attacks against data centres [and] this is not limited to physical attacks, but also cybersecurity attacks,” he said.
Mr Chinello also highlighted the importance of preventing and managing cyber threats. “However, we do not anticipate such geopolitical tensions to slow the ongoing trajectory of AI infrastructure buildout.”
