News

Public · Published

Big Tech Borrowing Isn't Crowding Out Treasurys, This CIO Says — WSJ

Debbie Cunningham, chief investment officer of Federated Hermes' $677 billion money-market funds, said borrowing by artificial-intelligence companies such as Alphabet, Microsoft and Meta need not crowd out demand for U.S. Treasurys, because companies hold the raised cash in money-market funds that buy Treasurys.

Published:

Updated:

What happened

Debbie Cunningham, chief investment officer of Federated Hermes' $677 billion money-market funds, said borrowing by artificial-intelligence companies such as Alphabet, Microsoft and Meta need not crowd out demand for U.S. Treasurys, because companies hold the raised cash in money-market funds that buy Treasurys.

Confirmed

Global impact / market context

If Cunningham is right, heavy AI borrowing does not push up Treasury yields by competing with government debt. Instead, cash raised sits in money-market funds that buy Treasurys, so company borrowing can support demand for government debt rather than reduce it.

Analyst inference

Some investors fear large AI-related borrowing raises Treasury yields, which means the interest rate the U.S. government pays. Cunningham says fund rules and collateral requirements leave many investors required to hold U.S. government debt, so that demand stays even as more lenders finance AI companies.

Analyst inference

What to watch

  1. Cunningham said hyperscalers like Alphabet, Microsoft and Meta do not spend all raised capital at once and park large amounts of cash in money-market funds that invest in Treasurys. Confirmed
  2. Watch whether Cunningham's view holds as AI companies keep borrowing: if their cash stays in money-market funds, Treasury demand should hold; if it is spent quickly, that support could shrink. Proposed
  3. Track fund rules and collateral requirements that oblige many investors to hold Treasury securities; changes to these rules could alter demand for U.S. government debt and the yields investors receive. Analyst inference

Evidence

  • Original source link unavailable