Jeff Gundlach: AI Chips as an Asset Class Could Be a Market Top Sign
Jeff Gundlach has criticized Nvidia’s plan with six asset managers to mobilize over $500 billion for AI infrastructure, questioning why assets of unknown lifespan are used as collateral for long-term debt. Mark Cuban compared AI chips as an asset class to crypto.
Gundlach argues that Wall Street has exposed its own vulnerabilities: "Assets of unknown life as collateral for long-term debt?" he wrote on X, adding that Nvidia's new financing consortium "will not likely age well."
He refers to a substantial arrangement where Nvidia partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to create platforms aimed at mobilizing over $500 billion so customers can purchase access to its chips.
His comparison is humorous yet laced with concern: issuing 30-year asset-backed securities against warehouses of bananas, even "newly engineered bananas of unknown life."
The core issue revolves around duration. The debt is long-term, while the collateral, a hardware class frequently replaced, has a shorter lifespan. Participants have acknowledged this: KKR described compute as critical infrastructure requiring "long-duration capital," and Jensen Huang argued that continuous CUDA software work "extends its useful life and improves its economics over time."
Gundlach's broader point is about pattern recognition. He urges investors to watch for proclamations of new asset classes built on financial engineering, supported by what he calls "questionable" credit ratings.
Others echo this sentiment. Mark Cuban predicted "chips as an asset class will be the new crypto," while Michael Burry has shorted AI companies, arguing that today's processors will soon become obsolete before spending is repaid.
The credit market has already shown signs of this concern, with Nvidia's own credit default swaps hitting a record after announcing $750 billion in AI deals. The Bank for International Settlements further warned that an AI bust could impact credit markets as severely as the 2008 financial crisis.
Gundlach's argument is more specific and responsive: not whether AI works, but whether a graphics processor retains value after seven years.