Tech giants have raised over $200 billion in AI bonds this year, but new issues keep breaking on debut.
Tech giants have raised over $200 billion in AI bonds this year, but new issues keep breaking on debut.

Tech giants have raised over $200 billion in AI bonds this year, but new issues keep breaking on debut.
Amazon, Alphabet, Nvidia, Meta, Oracle and SpaceX have raised over $200 billion in investment-grade bonds this year for AI infrastructure — nearly 15 times last year's $13 billion.
"The market can eventually absorb these volumes, but the scale and pace have caused 'indigestion,'" John Servidea, global co-head of investment-grade financing at JPMorgan, said.
New bonds from SpaceX, Amazon and Nvidia all broke below issue price in their first days of trading. Amazon's $25 billion offering saw its 10-year credit spread widen as much as 7 basis points within days, while Nvidia's comparable tranche widened about 5 basis points in the first week. Normally, roughly two-thirds of new investment-grade bonds see spreads tighten in the days after issuance, according to Bloomberg data.
The supply pressure is set to intensify. Morgan Stanley is arranging about $15 billion in debt for Anthropic's Texas data center project, and hyperscalers and data center operators are expected to bring another $50 billion to $60 billion of new financing after Labor Day. Goldman Sachs estimates roughly one-third of hyperscaler capital expenditure will be debt-financed in 2026, rising to 35 percent in 2027.
Underwriters are adjusting their approach. Some issuers now run non-deal roadshows before formally launching transactions to test investor demand and acceptable pricing ranges. Banks are also pulling large tech deals from weekly issuance forecasts to avoid telegraphing supply and prompting investors to reduce risk exposure early.
"An issuer preparing a $20 billion financing typically won't raise market expectations for weekly issuance volumes, because that would expose its participation and prompt investors to reduce risk exposure early," Mariya Entina, portfolio manager at DoubleLine, said.
For BlackRock's $12.5 billion financing for Meta's data center projects, underwriters JPMorgan and Morgan Stanley deliberately increased allocations to pension funds and insurers while reducing short-term trading accounts. The strategy worked — the new bonds' spreads tightened quickly after issuance — but it came at a cost: BlackRock had to offer a higher issuance premium to attract long-term capital.
Tech issuers are also managing market expectations more actively. Alphabet said its latest $25 billion bond sale would be its last US market financing this year, and it offered extra yield concessions to boost demand. The offering drew about $115 billion in orders, more than four times the issuance size, with the 40-year tranche's spread narrowing from an initial 155 basis points to 130 basis points.
Meta told investors after its $25 billion April financing that it would not return to the bond market until at least the fourth quarter. Oracle said in February it does not expect to issue again in 2026.
The supply pressure is also spreading to the high-yield market. Data center operator CoreWeave has shifted more toward leveraged loan financing, while Goldman Sachs is in talks with investors for about $5.4 billion to support BlackRock's QTS data center project tied to Microsoft.
Canada Life Asset Management fixed income fund manager Kshitij Sinha said investors are cautious: "Given every hyperscaler's capital expenditure plans, there's no urgency to add exposure right now."
The AI debt wave has implications beyond the bond market. Alphabet raised its 2026 capital expenditure target to as much as $205 billion, more than doubling last year's level, and posted its first quarterly negative free cash flow since its IPO. Fitch Ratings estimates AI-related investment added 1.4 percent to US GDP growth in the first quarter of 2026, and Morgan Stanley projects worldwide AI-related corporate debt issuance could reach $570 billion this year. If AI monetization falls short of expectations, borrowing costs for tech giants could rise, squeezing margins and pressuring equity valuations across the sector.
This article is for informational purposes only and does not constitute investment advice.