Hyperscale my debt!

PodcastUnhedgedEmily HerbertOct 1

Investors have lent AI-related groups about $500 billion so far this year, including roughly $200 billion to hyperscalers, and the speaker estimates hyperscaler borrowing could reach about $1 trillion by around 2030. Large bond deals are testing capacity in European and sterling markets, while speakers debate how corporate debt affects Treasury demand and highlight the difficulty of finding investments outside AI. The discussion also weighs the career risk of avoiding the dominant market theme against the risk that it falters, even as job gains point to a broader economy than the concentrated market.

AI Infrastructure Corporate Earnings Fixed Income Market Liquidity Investment Strategy Labor Market

Preview · 9 more points in the app

Read the Full Summary

Key Takeaways

Goldman Sachs said AI-related groups had borrowed about $500 billion so far this year, including about $200 billion from hyperscalers, and the speaker estimates hyperscalers could borrow about $1 trillion by around 2030. As uncertainty grows over future data-center CapEx and debt issuance, hyperscalers have had to pay more than governments and similarly rated companies to borrow.

Amazon's €14 billion bond and Alphabet's 100-year bond are testing or drawing attention to European and sterling corporate bond markets, while bankers said investors had absorbed billions of pounds of corporate bonds in one go. The speaker says large issues could encourage European companies to borrow at scale and issuers to use smaller currency markets, though some smaller markets may be nearing their capacity while the dollar market remains highly liquid.

The speakers describe AI exposure across stocks, private markets, and corporate bonds as making it harder to find relatively safe investments outside the theme, with KKR and the Bank of England cited as raising concentration concerns. Rob says investors who exit AI risk short-term underperformance while the theme dominates, even as one speaker describes the market as thin and concentrated despite job gains beyond AI.

MARKETS

Hyperscaler Borrowing — Goldman Sachs said investors had lent AI-related groups about $500 billion so far this year, including about $200 billion to hyperscalers. The speaker estimates hyperscalers could borrow about $1 trillion by around 2030.

“Yeah. So so far this year, Goldman Sachs said to us that investors have lent AI related groups about $500,000,000,000 The hyperscalers, they account for about 200,000,000,000 of that.”—

Borrowing Premiums — Later in the year, hyperscalers had to pay more to borrow than governments and similarly rated companies. The speaker attributes the growing premium to uncertainty about how much additional data-center CapEx—and therefore debt—these companies will issue.

“But later into this year, that has really blown out and they're having to now pay when they go to the market, especially in The US, quite a bit more…”—

European Bond Capacity — Amazon's €14 billion bond and other large hyperscaler issues are testing Europe's corporate bond market capacity. The speaker says these deals could show European companies they can borrow at scale and encourage issuance in smaller currency markets.

“Like, one is, as you mentioned, when they come to the market in euros, for example, Amazon comes along and says, boom, we're gonna issue like €14,000,000,000, like in one…”—

Hyperscalers Enter Bond Markets

The hosts introduce the surge in corporate borrowing and explain that hyperscalers are funding data-center expansion with debt, despite having historically relied on abundant cash.

AI Debt Reaches $1 Trillion

The discussion identifies the hyperscalers and estimates that AI-related groups have borrowed about $500 billion this year, with hyperscaler debt potentially reaching $1 trillion by 2030.

Big Tech Tests Global Bond Markets

Massive deals in euros, sterling, and other currencies are expanding smaller corporate bond markets, though hyperscaler issuance can crowd out other borrowers and push up borrowing costs.

Investor Capacity For Big Tech Debt

Investors have absorbed unexpectedly large corporate bond deals, but capacity may be more constrained in smaller markets, raising questions about shifting allocations among asset classes.

Hyperscaler Debt And Treasury Yields

The speakers debate whether hyperscaler bonds are diverting demand from Treasuries, with some investors seeing an effect while Rob argues that many Treasury buyers have fixed portfolio needs.

AI Concentration And Career Risk

The panel worries that AI dominates major asset classes and leaves fewer perceived safe alternatives, while investors weigh the career cost of avoiding AI against the risk of a reversal.

Soup, Accenture, And Sniffing

In Long Short, Emily goes long soup, Rob praises a listener’s successful Accenture pick, and Katie warns against sniffing mysterious items while cleaning a student’s room.