Tech companies' capital expenditure on AI, including R&D, is projected to reach $2.5 to $3 trillion annually. This figure, escalating from virtually zero a few years ago, is comparable to total global military spending and signifies a massive macroeconomic shift.
The concentration of software loan maturities in 2028 is not an impending cliff but a timeline for a market shakeout. Over the next few years, AI's impact will differentiate companies with durable business models that can refinance from those that are existentially threatened and will likely default.
The phenomenon of AI companies investing in and buying from each other is not a fraudulent bubble. It is a necessary market structure where capital-rich public firms provide attractive vendor financing to capital-poor private AI startups, enabling high-margin sales and fueling growth.
While gross debt issuance hits record highs, this is offset by a tremendous amount of maturities and coupon income. On a net basis, the market is growing slowly, revealing that the surge in AI-related financing is far more manageable than headline figures suggest.
Despite record issuance, tech bond spreads are not widening because hyperscalers are issuing exactly what the market craves: high-quality, long-duration debt. With rates at attractive levels, investors are eager to extend duration, creating a perfect supply-demand match that keeps the market stable.
Massive debt issuance by AI hyperscalers is fundamentally altering the U.S. investment-grade credit market. The tech sector's debt footprint is on track to exceed that of the entire U.S. banking sector, a significant structural change from the market's historical tilt towards financials.
To monitor systemic risk in the AI ecosystem, watch single-name Credit Default Swaps (CDS) for hyperscalers. Cross-asset investors use these liquid contracts to hedge a wide range of less liquid exposures like private debt and equity books, making them a key forward-looking risk indicator.