Goldman Sachs Forecasts Record Hyperscaler Bond Issuance Amid AI Spending Surge

What happened

Goldman Sachs published research projecting that five major hyperscalers-Amazon, Alphabet, Meta, Microsoft, and Oracle-will issue approximately $250 billion in bonds during 2026 and $400 billion in 2027 to finance artificial intelligence infrastructure investments. Through July 7, 2026, four of these companies had already issued $194 billion in bonds, a 79% increase from the roughly $108 billion issued in all of 2025. Goldman estimates hyperscaler capital expenditure will reach about $750 billion in 2026, against projected operating cash flow of roughly $778 billion, with debt issuance equivalent to approximately one-third of capital spending this year and around 35% in 2027. The firm noted that market saturation and issuer concentration are likely to become more binding constraints as issuance accelerates.

Why it matters

Goldman Sachs' projection highlights the unprecedented scale of debt financing underway in the technology sector as companies race to build AI infrastructure. The surge in bond supply is already affecting market dynamics: median borrowing spreads for 2- to 4-year bonds rose to 40 basis points from 30 basis points in 2025, while spreads on bonds maturing in more than 20 years increased to 118 basis points from 108.5 basis points. Cover ratios-a measure of investor demand-fell from nearly five times in February to below two times in July, according to Apollo Global Management, suggesting investors are becoming more selective. As a leading investment bank underwriting many of these bond sales, Goldman's forecast signals both the scale of near-term financing needs and potential pricing challenges ahead. The widening spreads and declining cover ratios indicate that future hyperscaler bond offerings may require more generous terms to attract sufficient investor interest.

Bigger picture

The hyperscaler debt boom reflects a broader shift in corporate finance as technology companies move from cash-rich balance sheets to significant external funding for capital-intensive AI projects. The combined U.S.-dollar debt footprint of these companies has more than doubled since September to over $360 billion, according to Sage Advisory. Secondary-market performance has deteriorated, with 78 of 91 hyperscaler bonds issued in 2026 trading at higher yields than at issuance, with a median increase of about 22 basis points. The median new-issue concession-the extra yield borrowers offer to price a bond sale-rose to 12 basis points in 2026 from 2.25 basis points in 2025. For investment banks like Goldman Sachs, this wave of issuance represents significant fee opportunities but also underwriting risk as investor appetite shows signs of fatigue. Goldman's analysis suggesting that wider spreads and falling cover ratios indicate investors are demanding greater compensation to absorb the supply points to a recalibration of risk pricing in the investment-grade corporate bond market.

What to watch

Monitor whether hyperscaler bond spreads continue to widen as issuance accelerates toward Goldman's $400 billion 2027 forecast. Track cover ratios and new-issue concessions on upcoming tech bond offerings to gauge investor appetite. Watch for any shift in hyperscaler capital allocation strategies if debt financing becomes prohibitively expensive. Keep an eye on Goldman Sachs' underwriting activity and fee income from technology sector bond issuance in upcoming quarterly results. Follow whether other investment banks revise their issuance forecasts or raise concerns about market capacity constraints.