Tech Bonds Yield 6%-8% as Hartnett Reverses His Short Call

By Olivia Hart

Technology-company bonds now offer yields of 6% to 8%, according to Michael Hartnett, Bank of America’s chief investment strategist, reversing his call from almost a year ago to short hyperscaler debt.

That reversal came after technology credit underperformed the broader asset class. Hartnett’s weekly “flow show” research note, published Friday, put the gap at 9% this year, the worst relative underperformance in three decades.

The larger yields now compete more effectively with stocks than they did before. Hartnett’s new stance rests on income and relative value, not on a claim that the sector’s capital needs have eased.

Why did Hartnett’s short call work?

Almost a year ago, Hartnett recommended shorting hyperscaler bonds. He expected a ballooning supply of debt in 2026 to force investors to demand wider spreads, making the existing bonds less valuable.

The issuance figures supported that supply argument. Hyperscaler debt issuance has surged to $182 billion so far this year, up from $17 billion raised just two years ago in 2024. That is nearly 11 times the earlier amount, and it helps explain why investors now require more for their capital.

Hartnett and his regular team—Anya Shelekhin, Myung-Jee Jung and Jessica Guo—used the same supply logic to interpret Meta’s financing. Their conclusion is not that technology companies have stopped needing capital. It is that creditors deserve a larger say in how that capital arrives.

His bearish call was right about supply; the open question is whether today’s yields compensate investors for the risks embedded in each issuer.

How did Meta’s equity deal protect bondholders?

Meta’s $95 billion equity financing this week gave Hartnett’s team a concrete example of their newfound respect for credit investors. The transaction used preferred shares and a new class of nonvoting C shares, placing more of the burden on equity investors.

The transaction did not issue new bonds, leaving existing bondholders out of the new financing burden. Hartnett viewed the deal as “credit-friendly” because Meta raised money through equity rather than adding to the company’s bond supply.

“Credit-friendly” should not be read as “risk-free.” The cited analysis does not set out the preferred-share terms, the use of proceeds or the full capital structure, so the structure protects bondholders in this transaction without settling the sector’s broader credit risks.

Meta protected bondholders in this deal by raising equity instead of issuing more debt.

Do 6%-8% yields make tech credit attractive?

Hartnett’s yield table shows why the new call is not a single bet on one company. Five-year debt from CoreWeave, which Hartnett grouped with riskier technology credit, carries a yield to maturity above 14%. Apple, which he described as more stable, is borrowing through 2062 at just over 6%.

The headline range therefore conceals a wide credit and maturity trade-off. The 9% relative underperformance also does not mean every technology bond lost 9%; it compares the group with the broader asset class. Investors still have to separate income from total return and assess each borrower’s credit.

The attractive case rests on competition with stocks, not equal treatment across technology issuers. CoreWeave’s 14%-plus yield comes with Hartnett’s description of greater credit risk, while Apple’s just-over-6% yield reflects a different borrower and a much longer maturity.

The 6%-8% headline is compelling, but the source’s own spread from just over 6% to more than 14% shows why a single yield figure would mislead.

What does the spending surge mean for buybacks?

Hartnett expects hyperscalers’ enormous planned capital outlays to crowd out open-market share repurchases. Those outlays are funded through both fixed-income and equity markets, and he concluded that the days of technology companies buying back their own shares are “probably over.”

The scale of the projected change is large. The sector repurchased $105 billion of its own shares in 2024. So far this year, that total is annualizing at just $14 billion, which is $91 billion below the 2024 level, or about 87% less if the pace holds.

The $14 billion figure is a run rate, not a final tally, so it cannot establish where the year will end. Even so, Hartnett’s financing analysis and the collapse in the repurchase run rate point in the same direction: more cash is likely to flow toward hyperscaler investment and away from buybacks.

If issuance keeps expanding while 6%-8% yields hold, tech bonds may deliver the competition with stocks Hartnett expects. If spreads widen again as supply builds, his original warning regains force. For investors, the practical test is whether that income outweighs the spread risk facing issuers after $182 billion of supply.