A surge in bond sales to fund AI investments is another factor pushing up bond yields.
Analysts point to the laws of supply and demand: if there is a jump in need for borrowing, lenders can charge higher interest rates, pushing up yields.
Five of the biggest AI hyperscalers — Alphabet, Amazon, Meta, Microsoft and Oracle — have issued $220 billion of debt already this year as they fund investments in data centres and models, LSEG data shows. This is more than double last year's total figure.
Borrowing for AI investments has helped push global corporate bond issuance to a record $4.9 trillion so far in 2026, LSEG data shows, up 14% from this point a year ago.
The U.S. Treasury recently announced bond buybacks which analysts say are aimed at limiting rising borrowing costs.
That initially helped stabilise the market, but long-dated bond yields have since crept back up.
Treasury Secretary Scott Bessent says that worries about rising debt and yields overlook the strength of the U.S. economy.
Central banks can also buy bonds if markets are stressed, as the Bank of England did during the 2022 UK mini-budget crisis.
The European Central Bank also has the power to buy government bonds to stem an "unwarranted, disorderly" rise in borrowing costs under its Transmission Protection Instrument, as long as a country facing stress complies with EU budget rules.
Many investors say the current rise in yields is orderly and reflects higher borrowing and inflation.
Falling oil prices would help short-term, but ultimately, longer-term borrowing costs will only come down durably once governments take concerted steps to bring down debt or boost growth, they say.
Unless they do that, bond vigilantes will be on alert.
The term refers to investors who seek to impose fiscal discipline on governments they perceive as profligate by demanding higher compensation to buy their bonds.
Investors can also demand more compensation if they think policymakers are failing to contain inflation.
TLT•
