Nvidia this week unveiled a plan backed by Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to commit up to $500 billion for AI data center construction. The more consequential element is Nvidia’s effort to build a secondary market for aging GPUs, a strategy that carries both risk and opportunity for the AI infrastructure market.
To attract institutional investors, Nvidia has agreed to guarantee with its own money that its chips used as collateral will retain value. If GPUs pledged as collateral fail to hold expected resale value, Nvidia will cover up to 25% of the difference between book value and actual market price after a default and liquidation. This is designed to address wrong-way risk, but it also means Nvidia’s obligations could increase exactly when demand weakens.
The announcement comes amid strain in traditional AI funding channels: Oracle has taken on significant debt, Google has issued new equity, and Meta has burned through cash. Microsoft CEO Satya Nadella even recommended the book 1873 during a recent earnings call, referencing railroad-era financial engineering that led to economic collapse. Nvidia has compared its approach to the Lucent Technologies story, but says it is not lending its own money—only backstopping a portion of collateral value while external investors shoulder most capital and risk.
Jensen Huang has described Nvidia’s AI servers as AI factories and long-term assets akin to railroads or airlines, not quickly depreciating PCs. He argues that a deep ecosystem of potential users—clouds, enterprises, researchers—can protect residual value and make older GPUs a cost-effective option for many AI workloads.
Meanwhile, a separate analysis estimates that major technology companies have accumulated about $570 billion in AI-related corporate debt as of early 2025, issued through investment-grade corporate bonds, convertible notes, and bank loans. The primary borrowers are hyperscale cloud providers and key suppliers racing to dominate AI computing. Despite the leverage, credit default swap spreads and corporate bond yields remain near historical lows because issuers have strong cash flows, debt is often secured by physical assets, and markets expect AI revenue to outpace spending.
Analysts warn that credit markets may be overlooking risks tied to uncertain ROIs, potential AI service price wars, and rapid hardware obsolescence. If new architectures emerge, the collateral backing some debt could depreciate sharply—something traditional credit rating models may not fully capture. A significant downgrade or default could reprice risk across technology credit and spill into broader markets, while the concentration of debt in a few systemically important companies adds vulnerability.
