AI Debt Boom Pushes US Investment-Grade Issuance Past 2020 Record
American companies have borrowed more in the investment-grade bond market this year than in any year on record, and the driving force is artificial intelligence. According to data cited in Chinese financial coverage, issuance is set to reach $1.76 trillion once Monday’s deals are priced, edging past the $1.75 trillion record set in 2020, when the pandemic sent companies scrambling for cash.
The comparison is striking. The 2020 borrowing wave was a defensive rush, as firms hoarded liquidity to survive lockdowns. The 2026 wave is an offensive one: companies are raising long-term debt to build data centres, buy chips and power the infrastructure behind large language models and other AI services. Many analysts had expected a record year precisely because AI spending has become so large.
The scale of that spending is now a market story in its own right. Citadel Securities, the market-making firm, argues that investors have placed so much capital behind AI that there is almost no room for disappointment. Nohshad Shah, who heads fixed-income sales for Europe, the Middle East and Africa at the firm, wrote that the recent sell-off in technology stocks shows investors are becoming more demanding about whether huge AI outlays will eventually produce enough profit.
With so much of the economy and the stock market riding on the AI trade, Shah wrote, other risks could surface, leaving the market with almost no margin for error. That is a notable warning from a trading house rather than an equity research desk, because it frames AI as a credit question as much as a stock-market one. Companies that borrow to build capacity must eventually service that debt, whatever happens to the share prices of the chipmakers and cloud providers leading the boom.
The borrowing has been helped by strong demand for corporate bonds, which has kept financing conditions friendly for large issuers. Investment-grade issuance records are usually a sign of confidence, but they also concentrate risk. If AI revenue growth slows, the same companies that rushed to sell debt could find refinancing harder or more expensive when their bonds mature.
For global investors, the message is that the AI build-out is increasingly financed with borrowed money rather than just equity and cash flow. That makes the corporate bond market a place to watch for early signs of stress. Spreads on AI-linked issuers, the pace of new deals in the final months of the year and commentary from rating agencies will all show whether lenders remain as enthusiastic as they have been so far.
The record also matters beyond the United States. US investment-grade debt is a benchmark for global corporate borrowing, and a shift in its pricing can ripple into Asian and European credit markets. For now, the appetite looks strong. The question is whether the returns from AI arrive fast enough to justify the debt taken on to build it.