Firms like OpenAI and Meta claim a compute shortage while also exploring selling compute capacity. This isn't a contradiction but a strategic evolution. They are buying all available supply to secure their own needs and then arbitraging the excess, effectively becoming smaller-scale cloud providers for AI.

OpenAI's strategy involves getting partners like Oracle and Microsoft to bear the immense balance sheet risk of building data centers and securing chips. OpenAI provides the demand catalyst but avoids the fixed asset downside, positioning itself to capture the majority of the upside while its partners become commodity compute providers.

While AI represents the largest segment of corporate debt, the risk is not yet systemic. The current build-out is primarily financed by the massive free cash flow from operations of megacap tech companies, not excessive leverage. The real danger emerges when this shifts to debt financing that cash flow cannot support.

Large tech companies are creating SPVs—separate legal entities—to build data centers. This strategy allows them to take on significant debt for AI infrastructure projects without that debt appearing on the parent company's balance sheet. This protects their pristine credit ratings, enabling them to borrow money more cheaply for other ventures.

The current AI infrastructure build-out is structurally safer than the late-90s telecom boom. Today's spending is driven by highly-rated, cash-rich hyperscalers, whereas the telecom boom was fueled by highly leveraged, barely investment-grade companies, creating a wider and safer distribution of risk today.

Unlike the dot-com or shale booms fueled by less stable companies, the current AI investment cycle is driven by corporations with exceptionally strong balance sheets. This financial resilience mitigates the risk of a credit crisis, even with massive capital expenditure and uncertain returns, allowing the cycle to run longer.

Zuckerberg compares the current AI build-out to historical infrastructure bubbles like railroads. He anticipates a potential collapse where over-leveraged companies fail, allowing well-capitalized firms like Meta to acquire valuable data center assets at a discount. It's a long-term strategic play, not just a fear.

Unlike past tech booms funded by venture capital, the next wave of AI investment will come from hyperscalers like Google and Meta leveraging their pristine balance sheets to take on massive corporate debt. Their capacity to raise capital this way dwarfs the entire VC ecosystem, enabling unprecedented spending.

Silver Lake cofounder Glenn Hutchins contrasts today's AI build-out with the speculative telecom boom. Unlike fiber optic networks built on hope, today's massive data centers are financed against long-term, pre-sold contracts with creditworthy counterparties like Microsoft. This "built-to-suit" model provides a stable commercial foundation.