Wall Street seeks $7.5T for AI buildout over five years

The numbers being thrown around for AI infrastructure spending have officially entered “too big to comprehend” territory. McKinsey projects $7 trillion in total data center investment by 2030, with roughly $5.2 trillion of that earmarked specifically for AI workloads. To put that in perspective, $7 trillion is more than the GDP of every country on Earth except the US and China.

And here’s the thing: most of this isn’t being funded with cash on hand. Debt instruments are doing the heavy lifting, making up about 75% of the financing model by some estimates. Credit spreads are widening, order books are thinning, and the financial plumbing that’s supposed to carry all this capital is starting to groan under the weight.

The hyperscaler spending spree

The companies writing the biggest checks are the usual suspects. Microsoft, Google, and Amazon are expected to spend between $660 billion and $700 billion on AI-related infrastructure in 2026 alone. Across the broader industry, that figure could exceed $1 trillion in a single year.

Alphabet and its peers have reportedly issued $159 billion in bonds to finance data center and AI investments in 2026, a massive jump from prior years. The bond market has essentially become an ATM for Big Tech’s AI ambitions, and the question is how much more the machine can dispense before something jams.

JPMorgan’s own estimates are slightly more conservative, pegging the necessary AI infrastructure spend at $5 trillion to $5.5 trillion.

Debt-fueled dreams and GPU-backed loans

CoreWeave, the GPU cloud provider, secured a $7.5 billion debt facility in 2024 backed by its GPU fleet. Repayments begin in 2026, and the interest rate sits around 11%.

That’s not cheap money. An 11% rate on $7.5 billion means CoreWeave needs to generate substantial revenue just to service the debt, let alone turn a profit.

Tech companies are increasingly turning to private credit and structured financial products, partly because regulatory constraints make traditional bank lending more complicated at this scale.

The overbuilding question

The bull case is straightforward. AI workloads are growing exponentially, enterprise adoption is accelerating, and every major company on Earth wants to integrate AI into its operations.

The bear case is equally compelling. Monetization timelines for many AI applications remain unclear. Most enterprises are experimenting with AI rather than deploying it at scale. And the compute efficiency of new models keeps improving, meaning you might need fewer GPUs tomorrow than you think you need today.

What this means for investors

But the second-order effects deserve attention. Widening credit spreads in tech-adjacent debt could ripple into broader fixed income markets. When $159 billion in new bonds hits the market from a single sector in a single year, it crowds out other issuers and pushes up borrowing costs across the board.

For crypto investors specifically, the dynamic is interesting. Despite the scale of AI investment, there’s a notable absence of cryptocurrency or blockchain-based tokens directly linked to this funding wave. The entire $7 trillion buildout is flowing through traditional financial rails: bonds, private credit, structured loans, and equity.