Morgan Stanley backs hyperscaler bonds as AI financing risks weigh on weaker issuers

Traders Union
All News
Financial Institutions
Morgan Stanley backs hyperscaler bonds

Highlights

Morgan Stanley favors high-quality hyperscaler bonds due to widening financing gaps and rising spread volatility for lower-quality issuers amid AI infrastructure expansion.
Since September 30, Microsoft, Meta Platforms, Oracle, and Alphabet issued $67 billion in new bonds across U.S. and European markets to fund AI-related growth.
AI company capital expenditures are forecast to reach $1.66 trillion in 2027, up 75% from 2026, despite potential project delays from local data center restrictions.

Credit outlook favors higher-quality issuers

As reported by CNBC, Morgan Stanley strategist Fernanda Lima says high-quality hyperscaler bonds are becoming more attractive as investors stay wary of the borrowing needed to fund AI infrastructure expansion.

She says the market is moving toward a wider gap in financing terms between investment-grade hyperscalers and lower-quality issuers. Hyperscalers have stronger balance sheets and ratings support, while junk-rated companies are typically more leveraged. Morgan Stanley expects improving risk-reward for high-quality hyperscalers, but more spread volatility for data center bonds and neocloud companies because of execution and funding risks.

Debt concerns have been building this year as AI-related issuance floods the bond market, pushing up yields for many names and widening five-year credit default swap spreads. Lower-grade bonds are under particular pressure, with CoreWeave emerging as a focal point after its bond yield rises above 16% earlier this month. The company eases some concerns after reporting third-quarter revenue above expectations and raising its full-year guidance, while its shares jump more than 25% on Wednesday.

Issuance pipeline and spending outlook

The market for hyperscaler debt remains open, Lima says, and recent transactions suggest these companies are willing to accept higher borrowing costs to keep funding expansion. Microsoft, Meta Platforms and Oracle have all issued new bonds in the U.S. since Sept. 30, while Alphabet taps the European market and some Meta bonds also go to Europe. Morgan Stanley puts total issuance across both markets at $67 billion over that period.

Lima says strong balance sheets and earnings cushions give hyperscalers the capacity to absorb higher financing costs while still pursuing attractive returns on AI investment. She adds that spreads have tightened somewhat for most names since recent deals were priced, with Oracle the main exception.

Morgan Stanley expects little additional issuance from AI companies for the rest of the year, but sees 2027 becoming much more active. The bank forecasts $1.66 trillion in capital expenditures from AI names next year, up 75% from 2026, and $1.77 trillion in 2028. Lima says project timing may shift as post-election conditions, including local voter support for data center restrictions, likely delay construction, but she expects hyperscalers to remain better positioned than other AI borrowers to secure power, start new projects and manage future uncertainty. She also says more transactions, including equity issuance, may be announced to help close the gap between investment needs and monetization.

The latest update follows our earlier article on AI data center financing risks amid U.S. project delays, rising costs, and power constraints.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.