Institutional
Bloomberg reports Wall Street underwriters are steering AI-linked hyperscaler bond deals toward pension and insurance buyers to stop fast-money investors from tanking the debt on its first day of trading
Since the start of 2025, Alphabet, Meta, Amazon and Oracle alone have issued more than $300 billion in bonds to fund AI infrastructure, but debt from Nvidia, SpaceX and Amazon has repeatedly cratered below issue price within hours of trading as fast-trading accounts flip the paper for quick profit. Underwriters are now favouring buy-to-hold real-money accounts over those investors and spacing out issuance, with bankers privately expecting $50-60 billion more hyperscaler debt to hit the market after Labor Day.
The mechanism here is worth sitting with because it is not really about AI credit quality, it is about who gets allocated the bonds. Real-money accounts, pension funds and insurers, buy to hold and do not care much what the bond does in the first week of secondary trading. Fast-money accounts, hedge funds and relative-value desks, buy specifically to flip, and when a $25 billion Nvidia or Amazon deal craters below issue price within hours, it is disproportionately their selling that does it. Underwriters reshaping allocation toward the buy-and-hold crowd is a direct admission that the AI debt market cannot yet absorb its own trading volume without visible price damage.
That is a tell about depth, not demand. Headline demand for hyperscaler AI bonds has been enormous, that is precisely why $300 billion has cleared since the start of 2025. But headline order books and secondary-market stability are different things, and the gap between them is where this story lives. A bond that is 5x oversubscribed at pricing and then trades below par by lunchtime tells you the order book was full of investors who never intended to hold it, which is a liquidity fragility problem dressed up as an oversubscription success story.
Spacing out issuance and steering allocation toward real money is the textbook fix, and it will probably work in the narrow sense of reducing day-one price cratering. But it does not change the underlying arithmetic: $50-60 billion more is coming after Labor Day, on top of an already-saturated market, from issuers whose AI capex still has to justify itself in earnings before any of this debt gets repaid from cash flow rather than refinancing. Pushing the paper into buy-and-hold hands reduces visible volatility, it does not reduce the amount of AI-linked leverage sitting on institutional balance sheets.
For anyone tracking AI-linked credit risk, the read-through is that underwriters are now actively managing optics around AI debt in a way they were not eighteen months ago, which is itself informative. When banks start curating who buys a bond specifically to avoid a bad first trading day, that is a market that knows it is thin and is trying to paper over it before a real stress event forces the question. Watch whether the pension and insurance buyers who are absorbing this allocation are doing so because the yield genuinely compensates for AI capex risk, or because underwriters made it hard to get allocation any other way.
Read the original: Bloomberg - Bloomberg reports Wall Street underwriters are steering AI-linked hyperscaler bond deals toward pension and insurance buyers to stop fast-money investors from tanking the debt on its first day of trading. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.