Big tech companies are currently driving the largest capital expenditure boom in modern history to build artificial intelligence infrastructure. However, a staggering $2.9 trillion of these commitments are not currently visible on their balance sheets.

In this episode of Risk vs Reward, our Editor at Research Bureau - Hari Viswanath explains the findings from a recent Portfolio deep dive into the hidden liabilities of the top seven AI hyperscalers. Through complex lease contracts, take-or-pay agreements for semiconductor chips, and variable interest entities, while companies are keeping massive infrastructure costs off their primary ledgers for now, they will eventually start reflecting in the financials of the companies one way or the other in the next few years .

With the era of easy money over and bond yields remaining elevated, this credit-fuelled expansion carries systemic risks. Hari details why this boom is fundamentally different from the dot-com bubble—which was primarily an equity event—and why an implosion in the AI infrastructure space could trigger a very real contagion effect across global markets.

Published on October 17, 2026