KBRA Releases Research – AI Credit: Off to the Races—AI Infrastructure Commitments Are Reshaping the Credit Profiles of the AI-7 Companies
26 Aug 2026 | New York
KBRA releases a report examining the credit implications of the rapid expansion of artificial intelligence (AI) infrastructure across the AI-7 companies—modern AI infrastructure leaders including Meta, Amazon, Alphabet, Microsoft, NVIDIA, Broadcom, and Oracle.
This KBRA report focuses on the significant growth in off-balance sheet commitments being used to secure future AI infrastructure capacity, including long-term leases, purchase and construction commitments, manufacturing and supply arrangements, cloud capacity agreements, and guarantees. Collectively, gross disclosed contractual commitments and maximum contingent support across the AI-7 have increased to approximately $3.2 trillion from $575 billion at the end of 2024. KBRA examines how these commitments are reshaping the companies’ broader financial risk profiles and reducing future financial flexibility despite generally strong balance sheets and substantial financial capacity.
Key Takeaways
- Contractual commitments have grown substantially. The AI-7 have approximately $3.2 trillion of gross disclosed contractual commitments and contingent support, significantly increasing the amount of future financial capacity already committed.
- Financial capacity remains strong. Six of the AI-7 maintain traditional lease-adjusted leverage below 1.5x, but rapidly growing contractual commitments are reducing future financial flexibility.
- Commitments are not economically uniform. Fixed and debt-like obligations, operating and capacity commitments, and contingent support should be analyzed separately, with outcomes driven by timing, utilization, demand, cancellability, and counterparty performance.
Click here to view the report.
Related Publications
- Data Centers in Focus: Credit Considerations Across Asset Classes Webinar Recap
- Data Centers: Developments and Trends in Project Finance
- Data Center Leases: Variations on Established Themes
- AI Compute Infrastructure: Cross-Sector Credit Considerations for GPU-, TPU-, and Accelerator-Intensive Assets
- AI Compute Financings: Evaluating Re-Leasing Risk for GPUs and TPUs