Ep. 037 - Who's Funding the $11 Trillion AI Buildout? (Capital Markets)

The global AI infrastructure build-out requires $11.3 trillion in cumulative capital expenditure through 2029, shifting the industry’s primary constraint from compute availability to financing. Hyperscalers and Nvidia are increasingly deploying "balance sheet as a service" models, where Nvidia provides residual value guarantees and off-take agreements to enable non-investment-grade entities to secure construction loans. This financial architecture is essential as AI debt financing surpasses traditional asset-backed markets like auto and student loans. Despite the rapid deployment of new hardware like the GB300, demand remains so robust that rental prices for older H100 GPUs continue to rise, driven by data center construction delays and the high profitability of frontier AI models. Lenders now face the challenge of evaluating execution risk and long-term asset viability in a market where traditional depreciation models fail to account for the sustained utility of high-performance compute clusters.