Jensen Huang, the founder of Nvidia, is attempting to transform the perception of the company's graphics processing units (GPUs) into long-term financial assets, akin to commercial real estate or toll roads.
This initiative is underscored by recent agreements with major asset managers like BlackRock and Goldman Sachs to create a $500 billion pipeline aimed at financing data centers and GPU clusters for companies that may not have the financial capacity to purchase these chips outright.
Huang argues that Nvidia's AI factory platform is a productive and revenue-generating asset that can maintain its value over time. However, the longevity and resale value of cutting-edge GPUs remain uncertain, as they typically depreciate after a few years.
Ben Emons, founder of FedWatch Advisors, highlights that the primary risk to this financing model is the potential for China to flood the market with low-cost silicon, which could drastically reduce the value of Nvidia's chips and the collateral backing the loans. Investors may view these GPUs as high-depreciation equipment, leading to demands for high-yield returns.
Despite these risks, Nvidia currently holds a dominant position in the U.S. AI chip market, with an estimated 75% market share, and rental rates for its H100 chips have increased significantly. The outcome of Huang's strategy and the broader AI infrastructure buildout will largely depend on how these dynamics evolve, particularly in relation to competition from China