The artificial intelligence (AI) boom still offers an investment opportunity, but Bank of America’s senior strategist Michael Hartnett believes the better deal now lies in lending to technology giants rather than owning their shares.
As of October 30th, he views selected Magnificent Seven bonds yielding 6% to 8% as increasingly attractive after a bruising selloff.
His argument turns the AI investment narrative on its head: companies spending aggressively to build tomorrow’s infrastructure could reward creditors even while testing shareholders’ patience.
The distinction matters as financing costs climb and cash demands intensify.
AI’s enormous funding bill changes the investment equation
The pressure is visible in Amazon’s (AMZN) latest earnings.
The company maintained its $220 billion capex forecast for 2026 while reporting $42.4 billion in negative free cash flow in the trailing 12 months.
Yet operating cash flow increased 29% to $168.3 billion.
That combination reveals the underlying problem: Amazon’s operations are producing substantial cash – but its infrastructure investments are consuming even more.
For shareholders, the “uncertainty” concerns when that spending will generate sufficient additional returns.
But bondholders face a different calculation: whether the borrower can meet interest payments and repay principal.
A business can remain creditworthy while delivering disappointing stock returns.
Hartnett’s preference therefore reflects a distinction between corporate financial strength and the price investors pay to participate in future growth.
Meta’s financing shift offers creditors a different signal
Hartnett’s figures put hyperscaler debt issuance at $182 billion this year, alongside $89 billion in equity issuance – a combined $271 billion.
That funding appetite helps explain why investors are demanding more compensation to finance the AI expansion.
Meta’s (META) plan to raise up to $95 billion through stock offerings illustrates how the consequences can differ across investors.
Additional common shares can dilute existing ownership, while equity financing can reduce reliance on further borrowing, potentially supporting creditors.
The package includes different securities, so its ultimate effect depends on the terms and execution.
Hartnett interprets the shift as evidence that the “era of big buybacks is over.”
That is his broader market thesis, rather than a declaration that every technology company will stop repurchasing shares.
Attractive yields still require careful bond selection
The appeal is straightforward: investors buying selected bonds at depressed prices can secure higher yields and potentially benefit if prices recover.
Unlike shareholders, they do not need expanding stock valuations to earn their contractual payments.
However, a quoted 6% to 8% yield is neither a universal offer across technology debt nor a guaranteed annual investment return.
Maturity, purchase price, seniority, and the issuing entity all matter.
Longer-dated securities can fall sharply when market interest rates rise, while deteriorating credit conditions can push prices lower even if Treasury yields decline.
Hartnett’s thesis ultimately rests on whether borrowing costs now compensate investors adequately for those risks.
For investors convinced that leading technology companies will remain financially resilient, their bonds offer a way to participate in the AI financing cycle with a different balance of income, upside, and downside exposure.
Comments
Log in or sign up to join the conversation.