(July 25): Big Tech’s insatiable appetite for artificial intelligence (AI) investment is unleashing a relentless wave of bond sales — and having a bigger impact on the US corporate bond market than might be obvious.
Debt from the biggest tech companies now has a bigger influence on market returns than bonds from the largest Wall Street banks, by one key measure. And more AI-related borrowing overall is coming: Alphabet Inc boosted its 2026 capital spending forecast, raising market fears of more debt sales, and a data centre project tied to Meta Platforms Inc is due to sell bonds next week.
The shift in market influence can be seen in a metric known as duration times spread, a measure of risk in a portfolio that looks at both credit risk and exposure to interest rates. The six biggest tech companies represent 8.6% of the DTS for the US high-grade corporate market, according to an analysis by Barclays strategists including Dominique Toublan and Andrew Keches, based on data as of July 23.
That’s more than the 7.3% DTS for the six largest banks in the Bloomberg US Corporate Bond Index, which historically have had much more sway in the market.
Investors have long taken solace in the fact that the biggest tech companies account for a relatively small part of outstanding debt in the market, with debt from hyperscalers including Alphabet, Amazon.com Inc, Oracle Corp, SpaceX and Meta comprising around 4% of principal in the market, including privately placed Rule 144A bonds. The biggest banks are about 9% of market value.
But the hyperscalers are in some sense riskier than the biggest banks, because they tend to borrow for longer periods of time. With their outsized bets on AI, some may be winners and some may be losers, or the whole industry may misjudge. Even if none of them default, betting on a company whose bonds weaken could result in a hit for a portfolio.
“The biggest risk isn’t credit quality, it’s concentration,” said John Fekete, head of tradeable credit at Crescent Capital. “If investors begin questioning the return on AI infrastructure spending, the resulting repricing could ripple through the entire bond market.”
That questioning has already begun: Alphabet’s shares fell after it projected up to US$205 billion (RM838.55 billion) of capex this year, and tech company corporate bonds and credit default swaps widened amid skepticism over whether the vast amount of debt-fuelled AI spending will justify returns.
With the likes of Meta, Microsoft Corp and Amazon all expected to detail their own AI spending paths in earnings in the coming week, fixed-income investors are bracing for an even larger wave of debt issuance.
Adding to the deluge of AI bonds in the market are roughly US$12.3 billion of bonds BlackRock Inc is selling to help finance a Meta data centre campus in El Paso, Texas. The debt tied to BlackRock’s stake in Project Sopaipilla Holdings has so far seen weak demand.
Credit markets are showing other signs of fatigue. Tech investment-grade bond spreads have widened, reaching about 89 basis points as of July 23, compared with around 76 basis points at the beginning of the year. And credit derivatives volume on the hyperscalers has jumped as investors and dealers look to shed risk.
When Amazon sold US$25 billion of bonds earlier this month, it got a cooler reception from investors. And when SpaceX, Elon Musk’s rocket, satellite and AI conglomerate, issued bonds in June, they weakened so quickly in the secondary market that traders said they couldn’t recall a recent deal that widened so sharply.
“As more AI-related debt is issued at spreads wide of the broad index, there is potential for this supply pressure to drag index spreads wider,” said Daniel Belton, a fixed income portfolio manager at Aegon Asset Management.
Barclays predicts that the companies known as hyperscalers will issue around US$285 billion of investment grade debt globally this year. Some investors welcome the tech-led issuance wave because the more highly-rated hyperscalers lift the market’s overall credit quality.
Long bonds
But as tech firms sell longer-dated bonds to lock-in capital while investor demand remains relatively robust, they’re commanding a greater share of the duration times spread, and thus having a greater impact on index risk than their debt footprint suggests. The biggest banks, on the other hand, tend to sell debt that matures within about 10 years.
“Whenever you have an industry gaining in share as a proportion of your universe, you have more concentration, and concentration is usually a negative,” said Mariya Entina, a portfolio manager at DoubleLine, who sees DTS as a key measure of risk when constructing a portfolio. “The next few years they’re set to issue more than they even issued this year, so I really don’t see that turning around for us anytime soon.”
Tech giants have issued more long-dated bonds relative to other high-grade companies. Over the past 12 months, about 43% of hyperscaler debt supply consisted of bonds with maturities exceeding 10 years, Bank of America analysts said in a note on Monday. By comparison, longer bonds made up 24% share of M&A debt deals and 23% of all other non-financial issuance.
Some of the long-dated hyperscaler debt is already trading at spreads close to BB-rated junk bonds, despite the fact that most hyperscalers command some of the highest ratings in the investment-grade category.
Still, the recent spread widening and cooling investor appetite are “rational responses” to the unprecedented wave of supply, said Shaun Caulfield, a research analyst at Lazard Asset Management.
“We’re going in with eyes open,” Caulfield said. “Workflows across our team today look completely different to what they did six months ago as a result of AI, and we know these tools demand enormous investment.”
Magnificent Seven
According to Barclays, the Big Tech debt mirrors the “Magnificent Seven” stock concentration starting around the 2010s era. However, while that was driven by rising stock valuations, today’s fixed-income concentration stems from heavy borrowing to fund AI infrastructure, the bank said in its report.
“On the current issuance trajectory, spillover to the broader index looks inevitable,” the Barclays strategists wrote, noting that as the market allocates more capital to tech, capacity for other issuers will shrink.
“That should eventually lift clearing concessions and secondary spreads beyond the hyperscaler complex,” they wrote. “At that point, concentrated supply pressure becomes an index-level problem.”
This comes as macro factors like Middle East geopolitics are already weighing on risk markets, forcing investors to simultaneously “layer in the sensitivities around what’s going on with investments in AI”, said Stephen Hooker, a portfolio manager at Newfleet Asset Management.
Doing so requires grappling with “the magnitude of the capex plans by the hyperscalers and the concerns over when and if they will yield a return on investment”, he said.
Uploaded by Chng Shear Lane