The AI Boom Is Built on Debt

The AI boom is often framed as a story of innovation, productivity, and limitless upside. However, beneath the hype lies a more uncomfortable truth that the AI revolution is increasingly being financed not by profits, but by debt and a staggering amount of it.

According to recent reporting from the Financial Times, the global race to build AI infrastructure is reshaping corporate finance. Tech giants are no longer just investing their own cash they are borrowing heavily, issuing bonds, and tapping private credit markets to fund what is becoming one of the most capital intensive buildouts in modern history. At its core, the problem is simple, the economics of AI don’t yet match the scale of investment. Companies are pouring hundreds of billions into data centers, chips, and energy infrastructure, but revenues from AI services remain far smaller. This gap is being filled with debt and lots of it which worries me.

This isn’t a small trend. Analysts estimate that AI related borrowing could reach into the trillions over the coming years, marking one of the largest debt cycles since the industrial era.

I've been talking about bubble and if this sounds familiar, it should. Periods of rapid technological change often come with financial excess. Railroads in the 19th century. Telecom infrastructure in the early 2000s. Housing before 2008. Each of these booms was fueled by debt, justified by transformative potential and followed by painful corrections when expectations outran reality. Friday we started to see that pain and what makes this cycle especially dangerous isn’t just the size of the debt it’s how it’s structured.

AI financing increasingly relies on layered, opaque arrangements involving private credit funds, securitized assets, and special purpose vehicles. These structures spread risk across a web of institutions, from banks to pension funds to insurers and potentially even ordinary retirement accounts.That interconnectedness means trouble in one corner of the system could ripple outward. A drop in AI demand, delays in data center construction, or declining value of hardware like GPUs could trigger broader financial stress. If you remember 2008 profits are privatized while losses are subsidized by us the tax payers. I hope our 401Ks and tax money don't get used if this goes south.