
Borrowing by major U.S. technology companies is rising sharply to fund AI infrastructure, while weaker cover ratios, wider concessions and softer secondary-market performance point to growing investor selectivity.
Major U.S. technology companies are selling debt at a much faster pace to finance artificial intelligence infrastructure, and investors are demanding higher compensation as supply builds. Amazon, Alphabet, Meta Platforms and Oracle issued about $194 billion of bonds in 2026 through July 7, up 79% from roughly $108 billion in all of 2025, according to a Reuters analysis of LSEG data. Goldman Sachs expects bond issuance by the five hyperscalers, including Microsoft Corp, to reach roughly $250 billion this year and $400 billion in 2027. The heavier issuance is pushing borrowing spreads wider across maturities. For Amazon, Alphabet, Meta and Oracle, the median spread on 2- to 4-year bonds rose to 40 basis points from 30 basis points in 2025, while 5- to 7-year debt widened to 60 basis points from 50 basis points and bonds maturing in more than 20 years widened to 118 basis points from 108.5 basis points. Secondary-market performance has also weakened: 78 of 91 hyperscaler bonds issued in 2026 with comparable pricing data were trading at higher yields on July 28 than at issuance, with a median increase of about 22 basis points. Demand remains strong in absolute terms but is weakening as investors absorb repeated jumbo deals. Apollo Global Management said cover ratios for hyperscaler bond sales fell from nearly five times in February to below two times in July, and research compiled by AlphaSense showed Amazon's March U.S.-dollar bond sale was about 3.4 times oversubscribed versus around 1.6 times for its July offering. Sage Advisory said Amazon's latest $25 billion bond sale pushed the spread on a 30-year Amazon bond issued earlier this year about 20 basis points wider. Reuters also found the median new-issue concession rose to 12 basis points in 2026 from 2.25 basis points in 2025. The pressure reflects the scale of AI spending. Goldman expects hyperscaler capital expenditure to reach about $750 billion in 2026 against projected operating cash flow of roughly $778 billion, with debt issuance equivalent to about one-third of capital spending this year and around 35% in 2027. Analysts say market saturation and issuer concentration could become more binding constraints if issuance keeps accelerating.