AI Infrastructure Spending Drives Surge in Corporate Bond Issuance
Morgan Stanley and PIMCO report a massive shift in AI investment from equities to the bond market as hyperscalers ramp up infrastructure spending.
The artificial intelligence investment boom has transitioned from an equity-driven narrative to a bond-market story, with AI-related borrowing now accounting for nearly half of year-to-date gross issuance of non-financial U.S. investment-grade debt with maturities exceeding 15 years. Morgan Stanley recommends investing in high-quality hyperscaler bonds, noting that firms such as Microsoft, Meta Platforms, Alphabet Inc., and Oracle Corporation have successfully issued $67 billion in new bonds since September 30.
Analysts have significantly revised capital expenditure forecasts upward for the five largest hyperscalers, projecting spending of $800 billion for 2026 and $1.1 trillion for 2027. If neocloud providers like CoreWeave and Nebius are included, spending could reach $1.6 trillion next year. This surge creates a positive outlook for semiconductor firms including Nvidia, Micron Technology, Broadcom Inc., and AMD, though it has also created a divide in financing terms. While hyperscalers remain well-capitalized, leveraged firms like CoreWeave have seen bond yields spike above 16%.
PIMCO projects AI hyperscaler debt issuance at approximately $250 billion this year, with the broader AI ecosystem potentially reaching $570 billion. Due to this concentration and potential construction delays following U.S. midterm elections, PIMCO suggests investors diversify into agency mortgage-backed securities and emerging market debt to mitigate risks associated with a potential build-out slowdown.
