J.P. Morgan estimates AI infrastructure ecosystem can raise $1.7 trillion additional debt, signaling substantial capital-raising capacity.

Major technology companies financing AI infrastructure retain substantial capacity to raise additional debt, according to J.P. Morgan analysis. The firm estimates that hyperscalers could collectively issue around $1.7 trillion of additional investment-grade debt before reaching levels that might trigger heightened investor concentration concerns.

Recent weakness in AI-related bond markets reflects investors reassessing pricing for new issuances rather than fundamental concerns about market capacity. "The current hyperscaler widening is a byproduct of the HG investor community trying to rationally price in an accelerating pace of issuance from the hyperscalers and not a reflection of insurers already hitting issuer risk limits for these issuers," J.P. Morgan stated.

Demand for hyperscaler and data centre financing is expected to remain robust, though pricing will likely remain dynamic as expectations around future funding requirements and monetization strategies evolve. As issuance has accelerated, investors have become more aggressive in dictating pricing and terms in primary debt markets.

Recent weeks have seen AI-related debt experience pressure, with investment-grade and high-yield high-performance computing bond indices widening sharply. However, J.P. Morgan characterized this as repricing driven by investor sentiment rather than any shortage of market capacity. The broader investment-grade market remains significantly less concentrated than during previous credit cycles, and insurers continue to have room to increase exposure to hyperscaler debt.

The AI and data centre financing universe now encompasses 31 issuers with more than $576 billion in outstanding bonds and over $5 billion in leveraged loans, and this investment universe is expected to continue expanding as AI infrastructure spending increases. Securitised data centre products have proven relatively resilient compared with corporate AI-related bonds, with the divergence between data centre project bonds and hyperscaler bonds reflecting an orderly repricing by investors rather than deterioration in underlying market fundamentals.