The highly rated corporate bond market is splitting between AI-related issuers and traditional companies as investors assess the scale and unpredictability of borrowing needed for data centers, chips and other AI infrastructure. Gross hyperscaler debt issuance is expected to reach a record $420 billion next year, 60% above 2026 estimates, according to Goldman Sachs data, while newer analysis shows hyperscaler debt up 92% year over year as capital expenditures rose 81%. Alphabet, Meta and Nvidia have faced wider spreads on larger offerings, although strong cloud and AI revenue growth and operating margins of 36% to 41% continue to support their credit profiles. The divergence reflects supply-and-demand pressure, concentration concerns and growing selectivity rather than a broad deterioration in AI borrowers’ credit quality.