Over the past year, the six largest American hyperscalers - Microsoft, Amazon, Google, Meta, Apple, and Nvidia - added $120B in debt. Cash on hand grew by just $58B. Net debt expanded from $236B to $298B - nearly a third in four quarters.
This isn't a coincidence or an accounting quirk. It's a direct consequence of the AI race that has been playing out for several quarters now.
Where the Debt Is Coming From
Two charts tell the story that leads here. First, AI capital expenditures began crowding out share buybacks - buybacks fell from $66B to $25B over the past year while capex climbed from $76B to $133B. Then, in Q1 2026, something happened for the first time in history: combined capex exceeded free cash flow - $133B against $119B.When spending outpaces what you earn, you have to find the difference somewhere. Big Tech is finding it in debt. This chart is the final piece of a single story about how the AI boom is actually being financed.
What the Numbers Show
In Q1 2025, combined debt across the six companies stood at $420B against $184B in cash - net debt around $236B. By Q1 2026, debt had climbed to $540B, cash moved up to $242B, but net debt had already reached $298B. The gap between what these companies owe and what they hold in cash has widened every single quarter.The scale matters here: $540B in combined debt across companies with trillion-dollar market caps isn't inherently alarming. The balance sheets can handle it, credit ratings are strong, and borrowing costs remain manageable. None of these players are under financial pressure in any traditional sense.
The issue is something else entirely.
Why It Still Matters
Debt is a future obligation. When a company borrows to build infrastructure, it's making a bet that the infrastructure will pay off. The larger the debt, the larger the bet - and the more expensive a miscalculation becomes.Big Tech has collectively borrowed hundreds of billions of dollars for AI, on top of what it has already spent from its own FCF and pulled away from buybacks. That's an enormous concentration of risk around a single theme. And the entire structure rests on one assumption: that AI monetization will eventually catch up to AI investment.
It hasn't yet. AI revenue is growing, but not at a pace that justifies the current scale of spending. The gap between what's being invested and what's being earned from AI remains very wide.
The Bottom Line
Three charts - buybacks vs capex, capex vs FCF, debt vs cash - tell the same story from different angles. Big Tech is reallocating capital at an unprecedented scale, funding AI infrastructure by pulling from shareholder returns, free cash flow, and now increasingly from borrowed money.That doesn't mean the bet is wrong. Two or three years from now, we may look back at these numbers as the most far-sighted capital allocation in American corporate history. But right now, the market sits at a point of maximum uncertainty: the money has been spent, the infrastructure is being built, and the returns are still ahead.
That's why the question of AI monetization timing is shifting from an abstract analytical exercise into something much more concrete - for market valuations, for balance sheets, and ultimately for everyone holding shares in these companies.