Nvidia (NASDAQ: NVDA) has signed memorandums of understanding with Apollo, BlackRock (NYSE: BLK), Blackstone (NYSE: BX), Brookfield (NYSE: BN), Goldman Sachs (NYSE: GS) and KKR (NYSE: KKR) to mobilise more than $500 billion in third-party capital for AI infrastructure.
The deal is drawing attention to a key issue: who carries the risk if AI hardware loses value faster than expected?
Under the proposed financing structure, Nvidia could guarantee up to 25% of the residual value of its chips in individual transactions. The arrangement is aimed at giving lenders greater confidence when financing customers’ purchases of Nvidia hardware.
Nigel Green, CEO of deVere Group, said the guarantee raises questions about the strength of demand for Nvidia’s chips.
“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?” Green said.
The comment does not suggest that Nvidia’s AI business is weakening. Instead, it highlights the financial risks created by the rapid pace of hardware development.
GPU depreciation becomes a financing issue
Nvidia releases new GPU architectures roughly every two to three years. That creates uncertainty for lenders financing hardware over longer periods.
A GPU that is considered high-value today could face stronger competition from newer architectures within a few years. Its resale value may then fall faster than lenders initially expected.
This matters because AI data centres require large investments in GPUs. Financing allows operators to spread those costs over time. But lenders must also estimate how much the equipment will be worth when the financing period ends.
Nvidia’s willingness to guarantee part of that future value could reduce some of that risk for lenders.
Green said the structure deserves scrutiny because similar asset-backed financing models have appeared in other sectors during periods of aggressive expansion.
Private credit takes a larger role
The financing plan also highlights the growing role of private credit in AI infrastructure.
Much of the capital is expected to move through private financing markets. These markets can provide large amounts of capital quickly. They also generally offer less public disclosure than traditional bank and public debt markets.
AI companies and data-centre operators are increasingly turning to such financing to fund GPUs, servers, power systems and other infrastructure.
For Nvidia, the model could help customers acquire more hardware without paying the full cost upfront. That could support faster expansion of AI computing capacity.
For investors, however, it makes the allocation of risk more important to understand.
If GPU values fall sharply, losses could potentially affect several parties, including equipment owners, lenders and investors. Nvidia’s guarantee could also expose the company to part of that downside.
Investors are watching sentiment
Green also pointed to a recent change in retail sentiment towards Nvidia. He said investor sentiment had shifted from bullish to neutral over the previous day, while online discussion volumes had fallen.
Such indicators can change quickly and do not necessarily reflect the views of institutional investors.
Still, the financing structure gives investors another issue to consider as they assess Nvidia and the wider AI investment cycle.
The central question is whether financing is simply helping meet genuine demand for AI infrastructure or whether increasingly complex financial arrangements are needed to sustain the pace of investment.
What the deal means for AI
The $500 billion figure is significant because it could help finance a large expansion of AI infrastructure. Nvidia’s GPUs remain central to many AI systems operated by cloud providers, technology companies and data-centre operators.
The financing could therefore support continued demand for Nvidia’s products.
At the same time, the structure highlights a broader risk. AI hardware is evolving rapidly, while the financing of that hardware can extend over much longer periods.
For investors, the key issue may be less about the headline value of the financing platform and more about how the risks are distributed.
Green said the arrangement should not be viewed as evidence that Nvidia is in financial trouble or that AI demand is not real. Instead, he urged investors to examine what happens if today’s hardware loses value faster than expected.
That question could become increasingly important as AI infrastructure investment moves from a technology race into a large-scale financing market.
