Nvidia and six Wall Street giants team up to raise $500bn for AI infrastructure

[Finley_27]

Nvidia has struck deals with six of Wall Street's largest financial institutions, Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR, to mobilise 500 billion dollars in third party capital specifically to fund artificial intelligence infrastructure, marking what several of the executives involved described as the first time AI compute itself is being treated as a genuine investable asset class

Nvidia CEO Jensen Huang framed the underlying logic plainly, saying in AI, compute is revenue, and that the company is bringing the world's leading long term capital providers together to independently underwrite AI infrastructure, he told CNBC he approached only these six firms for the commitment and none turned him down, the financing will use compute power itself as collateral, structured through private offerings and bonds issued by special purpose entities capable of raising tens of billions of dollars at a time, with Goldman Sachs, the only bank in the coalition, positioned to lead public debt deals while also distributing investment returns through its own asset management arm

The Wall Street executives involved offered genuinely striking framing of their own for why they're getting involved this way, Blackstone's Jon Gray said demand for AI is outstripping supply, with usage at Blackstone's own portfolio companies surging sevenfold this year alone, and compared AI compute to how mortgage lenders assess homes, a financeable asset class in its own right, BlackRock's Larry Fink went further, calling this the start of the next future for financial engineering, explicitly comparing it to the creation of mortgage backed securities back in the 1970s, and saying the US needs to raise this money as fast as possible because he believes it's imperative that America leads AI globally

KKR co-CEOs Joe Bae and Scott Nuttall offered a notably candid assessment of the actual challenge ahead, saying compute has become a critical infrastructure asset, and that as they've scaled their approach to digital infrastructure, they've learned that delivery, not ambition, is the hard part, the money raised will fund construction of new data centres to house, operate and cool the stacked computer chips that process AI workloads, as well as new factories to manufacture the chips themselves and increase their availability to buyers

This isn't happening in isolation either, essentially every major technology and AI company, Google, Meta, Amazon, Microsoft, SpaceX, Tesla, OpenAI and Anthropic among them, uses Nvidia's chips, and those companies have collectively spent over 1 trillion dollars in just three years on AI projects and infrastructure, demand that has driven Nvidia's own stock market value up fivefold over that same period, but not everyone is entirely convinced the enthusiasm is fully justified, Jane Sydenham, senior investment manager at Rathbones, told the BBC that while Nvidia clearly needs to keep facilitating AI's growth, the worry is that more and more money is going into these projects, and whether they're all going to earn the right return for the future remains an open question

Nvidia gets $500bn from Wall Street giants to develop AI projectswww.bbc.co.uk

[Emma92]

Compute is revenue is such a clean encapsulation of Nvidia's entire pitch here, framing chips not as a one time hardware sale but as an ongoing productive asset that generates returns is exactly the reframing needed to get pension funds and insurance capital genuinely comfortable underwriting this kind of infrastructure

[KingThomas62]

Larry Fink's comparison to mortgage backed securities in the 1970s is a loaded historical reference whether he intended it that way or not, that financial innovation famously enabled a huge amount of productive lending but also became the exact mechanism that amplified the 2008 financial crisis once the underlying assets turned out to be worth less than assumed

[DigitalNomad62]

KKR's delivery, not ambition, is the hard part quote is honestly the most grounded and useful thing said in this entire announcement, plenty of AI infrastructure plans exist purely on paper, actually building data centres, securing power and getting chips installed at this scale is a different and much harder problem

[Voyager17]

Jane Sydenham's skepticism about whether these projects will earn the right return is exactly the right question to keep asking as AI infrastructure spending keeps compounding, revenue from AI compute assumes sustained enterprise demand at prices that justify the capital cost, and that's genuinely not guaranteed to hold

[Arty Candle]

Using compute power itself as loan collateral is a fascinating and genuinely novel piece of financial engineering, the value of that collateral depends entirely on chips retaining their usefulness and market value over the loan term, and GPU generations become obsolete fast enough that this could be a riskier bet than it initially sounds

[Sorted Echo]

None of the six firms turning Huang down when he approached them shows either genuine conviction in the AI infrastructure thesis or a real fear of missing out on what's being framed as a defining moment in financial engineering, probably some combination of both depending on which specific firm you ask

[Barry]

Sevenfold surge in AI usage at Blackstone's own portfolio companies this year is a striking internal data point that adds real credibility to the demand outstripping supply argument, that's not speculative market sentiment, that's an actual asset manager watching real usage scale up across dozens of companies they own

[LurkingLegend]

The circular nature of these deals is worth watching closely, Nvidia has already signed hundreds of billions in deals across the AI ecosystem, and now it's also helping arrange the financing that lets its own customers buy more of its chips, that interconnected structure is efficient but also concentrates risk in a way that could amplify problems if AI demand ever genuinely slows

[Natalie_86]

$1 trillion already spent across just Google, Meta, Amazon, Microsoft, SpaceX, Tesla, OpenAI and Anthropic in three years, now layered with another $500 billion specifically being mobilized through Wall Street, shows the sheer scale this industry has reached, the real test over the next few years is whether the revenue these investments generate can actually justify capital deployment at this magnitude

[Wizard72]

This being explicitly framed as the first time AI hardware is being treated as an investable asset class is a significant structural shift for the whole industry, once institutional capital markets develop standardized ways to finance and trade exposure to compute infrastructure, that changes how quickly and how much capital can flow into the sector going forward