Nvidia's announcement this week that Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR are willing to commit up to $500 billion for AI data center construction has drawn significant attention—but the true headline lies elsewhere. Beneath the staggering figure is Nvidia's strategic push to create a secondary market for aging GPUs, a move with profound implications for startups and enterprises alike.
To win over these financial heavyweights, Nvidia has agreed to use its own capital to guarantee that its chips, when used as collateral in these deals, will retain their value. This assurance is central to the plan's appeal, yet it also introduces considerable risk.
A Novel Guarantee with Strings Attached
Specifically, Nvidia promises to cover up to 25% of any shortfall if GPUs used as collateral fail to hold their expected value. For instance, if a data center owner defaults on a loan and the lender is forced to liquidate, but the chips sell for less than the book value suggests, Nvidia steps in to bridge the gap. This mechanism is designed to provide confidence to lenders, but it carries a hidden danger: what financiers call 'wrong-way' risk. As demand weakens, Nvidia's obligations would grow—just as its own revenue faces downward pressure, creating a potential double blow.
Market Reaction and Clarification
The bond markets reacted nervously to the initial announcement, prompting CEO Jensen Huang to clarify the plan on X (formerly Twitter) and business television, emphasizing the limited scope of Nvidia's risk. This response was crucial, as many analysts have called the scheme 'unusual,' 'smart,' and 'dangerous'—all accurate descriptors.
Beyond Lucent: A Different Approach
Some have drawn parallels to Lucent Technologies, which collapsed after financing customers' purchases during the dotcom bubble. The comparison isn't entirely unfounded—Nvidia has committed billions to chip buyers, including frontier AI labs like OpenAI and Anthropic, and neoclouds such as CoreWeave (which pioneered using Nvidia chips as collateral), along with Nebius, Firmus, and Lambda. Bloomberg has even calculated that Nvidia has been working on a further $750 billion in circular deals this summer.
Yet Huang rejects the Lucent analogy. 'Is this circular financing?' he wrote on X. 'This initiative is designed to address that concern. We are bringing independent, long-term institutional capital into the AI infrastructure market.' And he's right—unlike Lucent, Nvidia is shifting most of the capital and risk onto external partners, merely by agreeing to protect a portion of its chips' future value.
A Secondary Market for AI Hardware
Underneath the financial maneuvering lies a more subtle, forward-looking agenda: Huang wants to ensure a thriving ecosystem for used AI hardware. By guaranteeing residual values, Nvidia aims to sustain demand for its products as they age, creating a virtuous cycle that benefits both the company and customers. For startups and enterprises, this could mean more accessible, affordable options in the secondhand market, extending the lifecycle of high-performance computing infrastructure.
In 2026, as AI workloads continue to evolve rapidly and data center costs soar, Nvidia's bold plan may prove to be a masterstroke—provided the risks are managed carefully. If successful, it could redefine how hardware companies finance large-scale deployments while fostering a more sustainable, circular economy for cutting-edge technology.
via TechCrunch AI