The AI Spending Spree: Tech Giants Bet Big With Borrowed Billions

The world's largest tech companies are making an expensive, and risky, pivot. To fund an unprecedented surge in artificial intelligence investments, firms like Amazon, Meta, and Alphabet are turning to the bond market in a big way, moving billions in speculative spending onto their once-pristine balance sheets. This shift is unsettling a core tenet for their investors.

For years, the understanding was that costly AI bets would be funded internally. "We've been told this spend would come from generated cash flow—that it was equity risk, not a credit concern," said Al Cattermole, a fixed income portfolio manager at Mirabaud Asset Management. "That unspoken contract is changing."

The numbers are staggering. UBS data suggests combined capital expenditure by AI hyperscalers could reach $770 billion in 2026, 23% above prior forecasts. To bridge the gap, these companies may borrow an extra $40-$50 billion this year. Oracle, Alphabet, and others have already launched massive debt offerings, including a rare century bond from Alphabet.

Suddenly, lenders are asking hard questions about creditworthiness for companies long seen as "cash-plus" entities. The concern is twofold. First, whether AI will generate profits fast enough to justify the debt. Second, that the very assets being built—vast data centers—could become obsolete by faster, more efficient chips, perhaps from emerging competitors.

"What if, in three years, these Nvidia chips get outstripped, and my data center is obsolete?" Cattermole posed, highlighting the long-term risk for debtholders. While firms like Vanguard note the sector's inherent strengths, they also warn of hidden risks building in the system through off-balance-sheet activity.

The message from the credit market is clear: the era of a blank check for AI is over. Every dollar borrowed now comes with an invoice, and a clock.