AI Debt Reaches 15% of Corporate Bonds, Displacing Banks

A surge in borrowing by technology companies to fund artificial intelligence infrastructure has pushed AI-related debt to 15% of the entire corporate bond universe, a historically high concentration that is displacing the world’s largest banks as the market’s dominant issuers. The shift creates a new form of systemic risk for fixed-income investors, mirroring the dynamic that has left the S&P 500 dependent on a handful of mega-cap technology stocks.

According to a recent analysis by Morningstar, the concentration of AI-related debt has reached a level with no historical precedent, creating a new risk profile for fixed-income investors. The scale of the capital required for the AI buildout is immense, with JPMorgan projecting that by 2027, investment in artificial intelligence will exceed all global military spending. This torrent of new issuance is concentrating risk further because it is happening as the broader corporate bond universe contracts, giving AI-related debt an even larger footprint.

The dynamic has produced a powerful, if narrow, rally. The S&P 500 has surged more than 16% over the past six weeks, an advance overwhelmingly driven by AI-centric names like Nvidia, Alphabet, and Amazon. Investors have so far been willing to look past rising bond yields and oil prices above $105 a barrel, betting that the productivity boom from AI can overpower macro headwinds. This has led to big tech firms tapping global debt markets, with Alphabet recently planning its first yen-denominated bond sale and Amazon preparing a debut in Swiss francs.

What is at stake is the introduction of a new, untested risk profile into a traditionally conservative asset class. While the credit ratings of a hyperscaler and a major bank may appear similar, their business models are fundamentally different. Banks borrow to lend at a higher rate, earning a spread. For tech companies, debt is being used to fund a massive capital expenditure cycle that must be serviced by future operating cash flows or uncertain AI-generated revenue, creating a potential single point of failure if the AI boom falters.

A Structural Shift From Banks to Tech

For decades, the largest financial institutions were the anchor tenants of the corporate bond market. That order is being fundamentally rewritten. The structural difference between these issuers is critical: for a bank, debt is a productive input in its core business of lending. For a technology company, that debt is a long-term bet on the future profitability of AI infrastructure like data centers, semiconductor chips, and the long-dated energy contracts needed to power them.

This distinction is becoming increasingly important as the AI debt pile grows. While the market is currently rewarding this investment, it introduces a vulnerability. The debt must be serviced from operating cash flows, which are dependent on the successful monetization of AI. Any downturn in the AI sector could now have an outsized impact on the corporate bond market, a risk that did not exist when bank debt dominated issuance.

The Political Economy of AI

Just as concentration risk in the bond market becomes apparent, a new political risk is emerging. In South Korea, the Kospi index recently swung violently after a presidential policymaker floated the idea of distributing AI profits to citizens in the form of an "AI dividend." The proposal, which would be funded by excess tax revenue from the AI boom, signaled that governments are beginning to view AI profits as a national resource requiring redistribution.

The market reaction was immediate and severe, with the Kospi falling more than 5% at one point as investors recalibrated the terminal value of AI profits. The incident, along with increasing demands from Samsung's labor unions for a greater share of the company's AI-driven profits, highlights a new vector of risk. The AI trade has been priced as a frictionless supercycle for corporations, but history shows that labor, and eventually governments, will demand a share of the spoils.

For bond investors, this means the assumption that AI profits belong solely to shareholders—and by extension, are solely available to service debt—is being challenged. The AI boom has so far traded like an endless expansion of market dominance. But the moment redistribution enters the narrative, the entire capital structure, including the mountain of new debt, must be re-evaluated.

This article is for informational purposes only and does not constitute investment advice.