By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panel expresses concern about the $500B financing of AI infrastructure, highlighting risks such as unproven securitization, maturity mismatch, and potential distortions to NVDA's earnings quality.
Risk: Mispricing or sudden liquidity stress due to unproven securitization and potential distortions to NVDA's earnings quality.
Opportunity: None explicitly stated.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
The first three-plus years of the artificial intelligence buildout has been paid for through record amounts of equity and debt issued by the world's leading tech companies, some of whom are spending so much of their existing capital that they've turned cash-flow negative.
Nvidia CEO Jensen Huang just revealed what he expects to be the next phase of financing, backed not by corporate balance sheets, but by Wall Street's top power brokers.
In an interview with CNBC on Monday, Huang called his plan a "big concept," unveiling it on camera alongside leaders from Goldman Sachs, BlackRock, Blackstone, KKR, Apollo and Brookfield. Together, those firms say they're willing to loan $500 billion, and potentially more, for the construction and buildout of new AI factories, as chipmakers and hyperscalers race to meet seemingly endless demand.
Huang and his big-money partners, one by one, described what they view as a fundamental shift in the tech industry: AI infrastructure has become a new asset class.
"These systems are not like our PCs, not like our phones," Huang told CNBC's Becky Quick. "These are revenue-generating assets now. They're productive, they're long lived, they're fungible, they're flexible."
The discussion was thin on specifics as far as the types of borrowers that will emerge, what interest rates will look like, where the facilities will be constructed and when it will all kick off. Their joint press release said the companies had signed memos of understanding, with no reference to any contracts.
The details matter. Almost 11 months ago, Nvidia announced a partnership to invest up to $100 billion in OpenAI as part of a plan to build out data centers requiring a combined 10 gigawatts of power. That investment never materialized, but Nvidia contributed $30 billion to the record-breaking funding round that OpenAI closed earlier this year.
Monday's announcement struck a different tone, with the companies collectively pushing the message that money won't be the problem as the AI buildout hits what McKinsey expects will be $7 trillion in global outlays by the end of the decade.
## 'These are real assets'
So far this year, Alphabet, Amazon, Meta, Microsoft and Oracle have raised well over $150 billion combined by selling debt and equity to build data centers and fund the development of new AI models and support the explosion of AI agents. Intel just announced a $15 billion stock offering, then upsized it to $20 billion.
Financial firms are now gearing up to jump into the market in a different way, as executives like Goldman Sachs CEO David Solomon and KKR's Waldemar Szlezak see AI equipment attaining familiar money-making characteristics.
"You're starting to see, in a sense, you know, asset-based financing against this infrastructure buildout," Solomon said on the CNBC panel. "That's not surprising because these are real assets. They have real value."
Instead of seeing supercomputers as devices that customers buy and use — the argument goes — these systems, filled with Nvidia's graphics processing units that can cost $3 million per rack, look like profitable investments. Huang says the systems can be improved through his company's CUDA software, and their lifespans extended, leading to better economics.
"You can think about it as a revenue stream, and you can securitize it or effectively divide that risk and sell it to investors who want to participate anywhere in that stack," said Szlezak, KKR's head of digital infrastructure.
When Wall Street starts getting noticeably excited about securitizing physical assets, a natural question emerges: What could go wrong?
One of the hallmarks of the financial crisis of 2007 to 2009 was the packaging of subprime mortgages into bundled securities that were then sold to investors as another way to make money from the housing boom. When mortgage defaults started going up, the whole system began to unwind.
Famed short-seller Michael Burry, who made a fortune betting against subprime mortgages, suggested late last year that companies including Meta, Oracle, Microsoft, Google and Amazon were overstating the useful life of their AI chips and understating depreciation.
The subprime meltdown wasn't part of the conversation on Monday, but several of the financiers acknowledged a certain amount of risk in the AI trade.
"There will be excesses, there will be pullbacks," said Jim Zelter, president of Apollo Global Management, adding that the number of participants in the project alleviates concentration concerns.
"There'll be big companies that win," Solomon said. "There'll be big companies that turn out to be not what people expected."
In discussing BlackRock's role in Monday's agreement, CEO Larry Fink made a direct comparison to the mortgage market, though he referenced a period decades before the housing boom and bust.
"This is the very beginning, like what it was when I started in the mortgage-backed securities market in the 1970s," Fink said. "I look upon this as as a next future for financial engineering."
All six of the financiers will make their own lending decisions, Huang said in the interview, noting that Nvidia will connect customers with financing partners.
Nvidia said it will have the option of backstopping 25% of every loan, a structure that should result in more favorable interest rates for companies that have previously had to rely on their own credit rating. Borrowers will have to use system architectures specified by Nvidia that would allow another company to take it over and operate it "if something were to happen," Huang said.
Nvidia still has plenty to iron out with its financing partners, but Monday's gathering marked a major step in showing the kind of money available to others in the ecosystem. Brookfield CEO Bruce Flatt said Huang created the necessary format for investors.
"Jensen's leading this to create structures," Flatt said. "Because there's hundreds of trillions of dollars of money in the world."
Four leading AI models discuss this article
"While the financing announcement removes a theoretical capital ceiling, concrete deployment hurdles around power, contracts, and true securitization readiness mean near-term revenue impact for NVDA remains uncertain."
Wall Street's $500B commitment to finance AI infrastructure as a new securitizable asset class is a logical evolution that could accelerate the buildout beyond the $7T McKinsey forecast by lowering capital costs for hyperscalers. Nvidia's 25% backstop and CUDA standardization reduce lender risk and extend asset lives, potentially supporting higher multiples on NVDA. However, the article glosses over execution risks: vague MOUs, no signed contracts, unproven securitization market for AI racks, and power constraints that could delay projects for years. Historical parallels to early MBS are seductive but ignore today's higher interest-rate environment and concentrated GPU supply.
This 'big concept' could simply be another round of hype that fails to materialize into actual loans at scale, much like the abandoned $100B Nvidia-OpenAI data-center plan, leaving balance sheets strained and exposing lenders to rapid AI obsolescence that Burry has already flagged.
"Nvidia is transitioning from a hardware vendor to a financial intermediary to artificially extend the AI capex cycle, masking potential demand saturation."
This $500 billion 'big concept' is a classic financial engineering play to sustain NVDA's parabolic growth by offloading capital expenditure risk onto private credit markets. By securitizing GPU clusters as 'real assets,' Huang is attempting to bypass the balance sheet constraints of hyperscalers. However, this creates a massive maturity mismatch risk. If these AI factories fail to generate the projected ROI, the collateral—specialized, rapidly depreciating hardware—will be worth a fraction of the debt. This move signals that organic demand growth is no longer sufficient to justify current valuations, forcing NVDA to become a de facto bank to keep its customers buying.
If AI infrastructure truly becomes a utility-like asset class, securitization could lower the cost of capital, accelerate global deployment, and cement NVDA’s hardware as the world’s standardized compute layer.
"Nvidia is quietly converting itself from a pure chipmaker into a quasi-financial institution bearing credit and obsolescence risk on assets whose underlying demand assumptions remain untested."
This is a confidence signal, not a financing solution. Wall Street announcing $500B in 'memos of understanding' for AI infrastructure is theater masquerading as commitment. The article buries the real issue: Nvidia is now backstopping 25% of loans, meaning it's absorbing credit risk on assets whose utility depends entirely on sustained AI ROI. If model economics deteriorate or demand softens, Nvidia becomes a shadow lender holding depreciating collateral. The comparison to 1970s mortgage securitization is revealing—not reassuring. We're watching the financialization of a technology whose cash flows remain unproven at scale.
If AI infrastructure truly generates predictable, long-lived cash flows (as Huang claims), then asset-based financing is rational and reduces balance-sheet strain on hyperscalers, freeing capital for R&D and accelerating the buildout cycle.
"The promised $500B of Wall Street financing is unlikely to materialize as stated and could unravel if demand or financing conditions deteriorate, creating liquidity and mispricing risks in AI infrastructure assets."
Initial read treats the $500B financing as a durable backbone for AI infra, but the article glosses over frictions. Memos of understanding are non-binding, and actual lending will hinge on credit quality, rates, and counterparty risk in a volatile funding environment. Data-center capex is energy-intensive, highly cyclical, and exposed to power costs and regulatory shifts. Securitizing these assets at scale is unproven, which could lead to mispricing or sudden liquidity stress if demand softens. Concentration risk among a small group of lenders and asset managers is real, and a downturn could trigger mark-to-market losses even with Nvidia’s 25% backstop.
The strongest bullish counterpoint is that if AI demand remains robust and rates stay favorable, securitized, long-duration capital could lower hurdle rates for data-center buildouts and accelerate AI adoption. This could create a durable funding moat for Nvidia's ecosystem.
"Nvidia's backstop creates hidden earnings distortion by shifting capex off hyperscaler balance sheets onto its own credit exposure."
Gemini's maturity mismatch claim understates the second-order effect: Nvidia's 25% backstop effectively turns it into a captive finance arm, subsidizing customer leverage while its own gross margins compress from credit provisions. Nobody has flagged how this distorts NVDA's reported earnings quality exactly as hyperscalers' capex rolls off their books.
"Nvidia is creating a circular credit loop that artificially inflates revenue while masking systemic credit risk through vendor financing."
Grok, your point on NVDA's earnings quality is critical. If NVDA is effectively subsidizing credit risk, we are seeing a 'vendor financing' trap disguised as infrastructure innovation. This inflates current revenue while masking future credit losses. I disagree with the panel's focus on the collateral; the real risk is the 'circularity' of capital. NVDA sells GPUs to hyperscalers using loans backed by NVDA itself. This isn't just financial engineering; it's a synthetic bubble waiting for a trigger.
"The circularity risk depends entirely on whether NVDA's 25% backstop remains contractually binding or becomes a moral hazard bailout."
Gemini and Grok are conflating two distinct risks. Yes, vendor financing distorts earnings—but the circularity argument assumes NVDA absorbs losses. If securitization actually transfers risk to private credit markets (the stated intent), NVDA's 25% backstop is a loss-leader, not a balance-sheet trap. The real question: does the 25% actually stick, or does reputational pressure force NVDA to cover more when assets underperform? That's where the synthetic bubble lives.
"The real test is whether securitizations truly absorb losses in a downturn or merely subsidize customers with Nvidia as a backstop."
Gemini, your 'circular capital' critique assumes securitizations genuinely transfer risk. The real test is whether the deals actually absorb losses in a downturn or merely subsidize customers with Nvidia as a backstop. If collateral values collapse (rapid GPU depreciation, duration vs. cash flows mismatch), losses could surface to NVDA or senior noteholders, risking marked-to-market stress and earnings quality well beyond what you imply.
The panel expresses concern about the $500B financing of AI infrastructure, highlighting risks such as unproven securitization, maturity mismatch, and potential distortions to NVDA's earnings quality.
None explicitly stated.
Mispricing or sudden liquidity stress due to unproven securitization and potential distortions to NVDA's earnings quality.