Citadel forecasts $500B in AI chip debt by 2028, reshaping credit markets
In brief
- Citadel Securities projects $500B+ in debt issuance through 2028 for AI chip financing
- Hyperscalers including Amazon, Microsoft, and Google have absorbed ~$570B in AI-related debt
- Debt maturity concentrated in 3-5 years, aligned with advanced AI chip lifecycles
- Chipmakers alone forecast $250B+ debt issuance; Anthropic secured $35B for Google processors
The Scale of AI Debt
Technology companies could issue more than $500 billion of debt by 2028, with most maturing within three to five years. The projected borrowing would equal more than 5% of the Bloomberg US high grade bond index by 2028, making it a structural shift in credit markets. Global markets have already absorbed about $570 billion of AI-related debt, much of it issued by hyperscalers including Amazon, Microsoft and Google.
The debt maturity profile tracks the hardware cycle itself. Most of the debt is expected to mature within three to five years, matching the relatively short useful life of advanced AI chips. This compressed repayment window creates refinancing risk and demand for continuous new issuance.
Chipmakers Lead the Borrowing Wave
Chipmakers alone could issue more than $250 billion in debt during 2028. AI developers face even steeper financing pressures. Anthropic secured a financing package of about $35 billion earlier in the year to purchase Google custom processors. Broadcom supported payments on the senior portions of Anthropic's financing transaction, allowing banks to distribute parts of the debt to investors.
Demand for external financing is rising among AI developers such as OpenAI and Anthropic as they spend heavily on computing capacity while generating negative cash flow. The pattern reveals a structural dependency: AI training requires massive upfront capital before revenue materializes.
Market Impact and Crypto Mining Implications
"The borrowing could develop into one of the largest new sectors in investment grade credit and significantly alter the composition of corporate bond portfolios." — Citadel Securities analysis
Citadel Securities expects the borrowing wave to influence credit spreads, portfolio construction and capital allocation across the broader AI industry. For crypto mining operations, tightening hardware availability and rising financing costs could reshape profitability models. Miners competing for GPU and ASIC inventory face steeper acquisition costs as tech companies bid up chip supplies.
Jeff Eason, Citadel Securities' head investment grade desk analyst, said the forecast may ultimately prove conservative as spending on AI infrastructure continues to accelerate. The debt issuance cycle will likely extend beyond 2028 if AI adoption and compute demand continue their current trajectory.
Frequently asked questions
Why is AI chip financing creating so much debt?
AI companies and chipmakers need massive upfront capital for data center infrastructure and chip production before generating revenue. Most of this debt matures in 3-5 years, matching the short useful life of advanced chips, forcing continuous refinancing and new issuance.
How does this affect crypto mining?
Tightening hardware availability and rising financing costs as tech companies bid for chip inventory could reshape mining profitability. Miners competing for GPUs and ASICs face steeper acquisition costs in a supply-constrained market.
