AI infrastructure credit continues to offer attractive investment opportunities, but the latest results from Amazon, Google, and Microsoft suggest those opportunities are becoming more nuanced. For investors, greater scrutiny will be needed. Key takeaways: Q1 results from the world’s biggest AI infrastructure hyperscaler companies are painting a compelling market picture – not only is AI demand continuing to grow, but it is also driving broadening monetization and growing order backlogs. However, dig beneath the headlines, and it is clear that AI capacity is struggling to keep up, raising the risk that capex shifts from cyclical build-out to a more structural base requirement and becomes a recurring cost factor in hyperscalers’ models. With capex-led issuance expected to increase in 2026 and beyond, greater investment scrutiny will be needed: we favour tactically positioning in those semiconductor, memory, and full-stack hyperscaler issuers that can benefit from capacity scarcity, with balance sheet strength and durable pricing power