AI’s Debt Surge: Wall Street’s New Money Trail

The push to build AI infrastructure is turning into a huge borrowing wave, says JPMorgan. The bank now thinks that by 2030 the world will have issued about $4. 1 trillion in AI‑related loans, a jump from earlier estimates. This surge comes as companies that grow data centers and buy chips scramble for cash. In 2026, AI borrowing already topped $300 billion. Data‑center lenders are leading this trend, and the bank expects AI spending to climb to $5. 5 trillion by 2030. That figure is up from a previous forecast of $5. 1 trillion and reflects a predicted rise in data‑center capacity to 138 gigawatts, up from an earlier 122 gigawatts. Developers are turning to creative power solutions—behind‑the‑meter deals, bring‑your‑own‑power options and more efficient hardware—to keep costs down.Despite the scale of spending, tech giants are still in a good position to pay for expansion. JPMorgan projects that hyperscaler companies will spend $650 billion in 2026 and over $1. 1 trillion in 2027, with operating cash flow potentially exceeding $900 billion that same year. The focus on credit markets is growing. While investors have mainly looked at chipmakers and software firms, JPMorgan believes that high‑grade corporate debt will fund more than $2. 1 trillion of AI infrastructure over the next five years, with another $350 billion coming from leveraged finance. Loan‑to‑cost ratios are high, often above 85% of project costs. The next big funding hurdle may be the chips themselves. The bank estimates that more than $3 trillion will be needed to buy GPUs and custom AI accelerators in the next five years. Private lenders are stepping up, but public markets will likely need to step in with larger sums. In short, the AI boom is not just a tech story—it’s becoming one of Wall Street’s biggest financing tales, driven increasingly by debt.