Morgan Stanley sees $2.9 trillion in data center spending, and a $1.5 trillion hole

The spending breakdown shows $1.6 trillion allocated to IT hardware including chips and servers, while $1.3 trillion is earmarked for physical infrastructure and construction. Major hyperscalers like Amazon, Microsoft, Alphabet, Meta, and Oracle are projected to self-fund approximately $1.4 trillion through their own cash flows, leaving a significant external financing gap.

Financing Gap

Morgan Stanley's research indicates private credit markets are expected to contribute roughly $800 billion toward the $1.5 trillion shortfall through asset-specific structures where loans are tied to particular facilities or equipment rather than company balance sheets.

Hyperscaler lease commitments already exceed $1.2 trillion in undiscounted future payments, which support much of the construction lending. Any pullback in hyperscaler demand would significantly impact developers and their creditors.

Power Constraints

Financial constraints aren't the only challenge. Revised forecasts show the US data center power shortfall increasing to 32-33 GW by 2028, representing a physical limitation that cannot be financed away.

Nvidia has been involved in discussions over backstop guarantees potentially reaching $250 billion to support data center debt financing, though these talks remain unconfirmed.

Government Intervention

Recent analysis suggests private credit is funding the AI buildout against collateral with a short clock, prompting speculation about potential government intervention. Some market observers believe Washington may eventually backstop the private financing used for AI data centers, particularly as risks accumulate.