Karen Finerman on Hyperscaler Debt and AI Funding

Hyperscaler Debt Issuance

• 💡 Hyperscalers are increasingly tapping the bond market to fund their AI ambitions, with companies like Meta and Oracle raising over $75 billion since September.
• 📈 The market's comfort level with these companies is reflected in their bond spreads, which have widened, indicating it's becoming more expensive to borrow.
• ⚠️ Spreads have widened significantly in the last month, suggesting a creeping fear about the promised returns on these AI investments.

Market Dynamics and Borrowing Costs

• 💰 Bond buyers are demanding higher compensation to offset the perceived risk, leading to increased borrowing costs for hyperscalers.
• 📉 The sheer volume of debt expected to come to market is likely to cause spreads to widen further.

Apple's Unique Position

• 🍎 Apple stands out as its bond spreads have remained nearly identical and haven't moved, suggesting it's perceived as having US government credit quality.
• 🔄 This contrasts with other hyperscalers facing a vicious cycle of needing to pay more to borrow.
• 💡 Apple previously executed smart debt strategies when rates were near zero, selling debt to buy back stock, creating a virtuous cycle.

Future Supply and Funding

• 🚀 Significant new bond issuance is anticipated, especially if companies like Meta plan substantial capex increases next year.
• 💬 While Oracle has significant capacity for build-out, the market's reaction to Meta's capex comments highlights concerns.
• 🏦 The discussion touches on the idea of a government backstop for entities like OpenAI, though this is speculative.