US investment-grade companies have sold a record $1.88 trillion of bonds so far this year, according to data from LSEG through November 4. That's up 18.7% from the same period last year and already tops the previous annual record of $1.86 trillion set in 2020.
The headline number is impressive, but the story behind it is less about a flood of new borrowers and more about a handful of tech giants borrowing enormous sums to fund artificial intelligence infrastructure. Amazon and Alphabet led the charge, and together with Meta, Microsoft, and Oracle—collectively known as "hyperscalers"—they've raised more than $370.7 billion, or roughly one-fifth of all US investment-grade supply.
Jumbo deals, fewer of them
One striking detail: the number of deals actually fell about 4% to 1,084, yet total volume still hit a record. That means the average deal size has grown significantly. Companies aren't just tapping the market more often—they're borrowing bigger chunks each time.
Apple added to the wave with a roughly $7.5 billion multi-tranche deal, and buyers have largely shown up. Crossmark Global Investments, an asset manager, noted that offerings in both US dollar and euro markets have been oversubscribed, meaning demand exceeded supply.
But there's a catch. When supply is this heavy and concentrated, investors often demand a little extra yield on new bonds to make room in their portfolios—even when order books look full. Underwriters may have to offer a "new-issue concession," a small premium over similar existing bonds, to entice buyers to reshuffle their holdings.
AI uncertainty shows up in credit insurance costs
Some investors are also using credit default swaps (CDS)—essentially insurance against a company defaulting—to price in the risks of AI's massive capital spending. Reuters noted that Oracle's five-year CDS spread hit a record 325.45 basis points last month. A basis point is one-hundredth of a percentage point, so that's 3.25 percentage points of annual insurance cost.
That record CDS level is a stress test for the entire $370.7 billion of hyperscaler bonds. When a handful of repeat issuers make up around a fifth of the market, they can influence the going rate for everyone else. Higher CDS spreads raise the implied cost of hedging and can foreshadow wider credit spreads—the extra yield investors demand over US Treasuries—in investment-grade tech.
What it means for investors
For everyday investors, this matters in a few ways. First, if you own bond funds or ETFs that hold investment-grade corporate debt, the record supply could mean slightly higher yields on new issues, but also potentially wider spreads on existing bonds. That can affect the value of your holdings.
Second, the concentration of supply means Big Tech's borrowing habits can ripple through the entire market. When these giants dominate issuance windows, other companies trying to sell bonds may face tougher pricing or have to wait for a quieter moment. That could translate into higher coupons on future mega-deals and more expensive borrowing for everyone else.
Finally, the record CDS spread on Oracle is a reminder that even the biggest tech names carry risk. AI spending is a bet on future growth, and if those bets don't pay off, bondholders could feel the pinch. It's not a reason to panic, but it's worth watching how credit markets price AI risk in the coming months.
For context, this isn't the first time corporate debt markets have seen a surge. In 2020, companies borrowed heavily to shore up liquidity during the pandemic. This year's record is different—it's driven by growth ambitions rather than crisis management. That's a healthier sign, but it also means the debt is being used to fund long-term projects with uncertain payoffs.
As the year winds down, investors will be watching whether the pace of issuance continues and whether credit spreads stay stable. If Big Tech keeps borrowing at this clip, the record could grow even larger. But the bigger question is whether the AI boom justifies the debt—and that's a story that will unfold over years, not months.
