Goldman Sachs reports that its high-yield AI datacenter credit basket is currently trading at a spread of 353 basis points. This figure exceeds the levels recorded in June 2022. The bank launched this instrument in July 2026, initially comprising 18 equal-weighted US high-yield companies with an average spread of 319 basis points and a yield of 7.45%.
Broader AI-related bond issuances for 2025 and 2026 show an option-adjusted spread of 381 basis points. Goldman Sachs estimates that total AI-related debt issuance will reach nearly $500 billion in 2026, accounting for approximately 18% of the total US investment-grade supply. Market data indicates that 17 out of 23 tracked datacenter joint ventures are trading wider than their original yields.
The bank's separate AI leadership basket, which tracks investment-grade credits, saw its spread expand from 74 basis points to nearly 150 basis points over the last 12 months. Additionally, new-issue concessions for large deals have increased by up to 20 basis points due to cooling demand for longer-dated tranches.
Rising spreads lead to higher borrowing costs for independent datacenter operators and joint ventures. Goldman Sachs notes that an additional spread of 30–40 basis points on hundreds of billions in debt results in billions of dollars of incremental interest expenses. In response, the bank has introduced custom baskets and swaps to allow investors to trade AI credit exposure.
