Investors harbor a veritable obsession with hyper-scaler capex and for good reason.

As discussed here and everywhere else ad nauseam, it’s a big deal when the companies which together comprise more than a third of large-cap US equity market cap transition from capital-light to capital-intensive businesses virtually overnight.

Although we’re talking about some of the best-run enterprises in the history of capitalism, the fact is, these management teams in some cases don’t have a lot of experience running businesses which require massive capital expenditures, nor businesses with highly-levered balance sheets.

Whether these concerns are overblown’s a matter of opinion — and also a matter of what metric you consult while conjuring the ghost of the dot-com bust, an unavoidable analogue particularly given the similarities between the web of deals at the heart of the AI boom and TMT-era vendor financing.

With that in mind, I have some comforting news: The run-up in tech debt burdens from the launch of ChatGPT through 2026 looks nothing like the 1996-2000 tech debt binge, at least not in percentage terms.

Indeed, the figure on the left, below, from Goldman, suggests this is apples to oranges. “The roughly 10% increase in TMT-related debt since 2022 pales in comparison to the rate of growth during the dot-com boom,” the bank’s equities team remarked.

The figure on the right gives you some context for the $1.67 trillion in total (i.e., IG plus HY) US corporate debt issuance over the past four quarters: It’s nearly a record in absolute terms, but as a share of assets, it’s quite tame.

Similarly, the S&P’s trailing net debt/EBITDA ratio is just 1.6x, middling on a 45-year lookback, and as Goldman’s Ben Snider observed, the hyper-scalers “could add $700 billion of debt and still carry a net debt/EBITDA ratio below 1x.”

Still — and this brings us full circle — we can’t just overlook the fact that the biggest companies in the world are set to become heavier borrowers. Goldman’s credit team noted that a third of all USD net debt supply last year was AI-related, and AI capex will play a big part in what’s expected to be $2.3 trillion in total IG and HY supply during 2026.

One thing’s for sure: That flood of supply bodes well for Wall Street investment bankers. Just ask Morgan Stanley’s Q4 debt underwriting haul.