JPMorgan: AI Spending to Drive Record U.S. Investment-Grade Bond Supply by 2027
nashnova research
JPMorgan forecasts U.S. investment-grade bond issuance will hit $2.1 trillion in 2027, with AI infrastructure as the core driver — the debt market's "new normal" is being rewritten by tech capex.
Will $2 trillion become the new normal?
JPMorgan strategist Nathaniel Rosenbaum and colleagues projected in a November 14 report that 2027 U.S. investment-grade bond issuance — debt from highly rated, lower-risk corporate borrowers — will reach $2.1 trillion.
This year's expected volume is already a record $2 trillion; next year adds another 5%.
The report states outright: annual issuance of $2 trillion or more "may increasingly become the norm." This means → the volume of new debt the market must absorb is stepping up permanently.
Where is the money going?
Tech and media companies are expected to raise a combined $612 billion, roughly 30% of total investment-grade supply.
Hyperscalers — Amazon AWS, Microsoft Azure, Google Cloud and peers — account for $225 billion; data-center and compute-chip financing adds another $225 billion.
In plain terms = for every three dollars of investment-grade debt issued, nearly one dollar funds AI infrastructure.
Rosenbaum noted supply will be "unusually concentrated" in AI-infrastructure sectors. This reflects an extreme sensitivity: the entire forecast hinges on the AI capex cycle continuing.
M&A debt is shrinking — but are "animal spirits" dead?
M&A-related bond supply is expected to fall 52% year-over-year — more than halved.
Yet strategists flag that stronger "animal spirits" — the risk appetite of market participants — remain the primary upside risk to both M&A and overall issuance.
This means → even as deal-driven financing contracts, a shift in market sentiment could push actual issuance above forecast.
What could hold things back?
Political risk: Democratic gains in the midterm elections may trigger permitting and development delays for data-center projects in key states, suppressing tech-sector borrowing demand into 2028.
Funding diversion: if hyperscalers continue shifting more debt issuance into non-dollar currencies, dollar-denominated supply takes a hit.
In plain terms = if data centers can't get built, or if Big Tech borrows elsewhere, the $2.1 trillion figure shrinks.
Can Wall Street even agree on a number?
TD Securities forecasts 2027 issuance surging to $2.4 trillion — more bullish than JPMorgan.
Bank of America sees volume dropping back to $1.9 trillion, citing hyperscaler funding diversification and cooling M&A activity.
The three forecasts span $1.9 trillion to $2.4 trillion — a $500 billion gap. This reflects a market that has no answer to one core question: how long can the AI capex cycle last?
市场有风险，内容仅供研究参考，不构成投资建议。