Amazon Secures $17.5B Credit Facility for AI Investments

Amazon.com entered into a $17.5 billion senior unsecured delayed draw term loan (DDTL) credit facility on June 8, 2026, with Citibank N.A. as administrative agent, per an SEC Form 8-K filing reported by Quartz and Reuters. The facility, which carries no financial covenants, allows Amazon to draw funds as needed until September 30; any borrowed amounts mature three years from the draw date. Amazon stated the funds are for general corporate purposes, while Bloomberg, cited by PYMNTS, described those as including AI-related capital expenditures. The loan arrives as Amazon pursues one of the largest capital spending programs in corporate history: $200 billion in 2026 capex, with Q1 spending of $44.2 billion up from $25 billion a year earlier.
What happened
Amazon.com entered into a $17.5 billion senior unsecured delayed draw term loan (DDTL) credit facility on June 8, 2026, with Citibank N.A. as administrative agent, per an SEC Form 8-K filing covered by Quartz and Reuters. The facility allows Amazon to draw funds as needed rather than taking the full amount at once; commitments expire September 30, 2026, and any borrowed amounts carry a three-year maturity from the date of each draw. Amazon stated that borrowing "will be used for general corporate purposes." Bloomberg, cited by PYMNTS, described those purposes as including "supporting business investments, funding future capital expenditures, and repaying debt."
Facility terms
The loan carries no financial covenants and allows flexible prepayment without premium, per Quartz reporting on the filing. Interest is calculated at Amazon's option using either a floating base rate (zero margin) or a term SOFR rate plus a margin of 0.625-0.875%, depending on the company's credit ratings. The participating lender group includes institutions that already provide a range of banking and capital markets services to Amazon.
Capital spending context
Amazon disclosed alongside Q4 2025 earnings that it planned roughly $200 billion in capital expenditures for 2026 - the bulk tied to AI infrastructure including data centers and custom chips - per Quartz. Capital expenditures reached $44.2 billion in Q1 2026, up from $25 billion in Q1 2025. On a trailing 12-month basis, free cash flow fell to $1.2 billion from $25.9 billion, reflecting a $59.3 billion year-over-year increase in property and equipment purchases. Amazon CEO Andy Jassy, speaking to CNBC in May 2026, defended the spending scale: "When you have shifts that are this momentous ... you want to bet big." He drew a parallel to Amazon's early AWS investment, arguing the pattern of heavy upfront capital spending eventually producing strong margins would repeat at a larger scale with AI.
AI context
Amazon's AI business had reached an annualized revenue run rate above $15 billion as of Jassy's CNBC comments in May, per Quartz. The loan is part of a broader trend in which large cloud providers have become major borrowers in the U.S. investment-grade corporate bond market as AI-related capital spending surges. Amazon also sold approximately $10 billion (14 billion Canadian dollars) of Canadian-dollar high-grade bonds on June 8, in parallel with the term loan agreement.
What to watch
Whether Amazon draws on the DDTL Facility before the September 30 commitment expiry, and any earnings-call language that clarifies use of proceeds beyond the general corporate purpose language. AWS capacity announcements, new instance types, or chip procurement partnerships would signal how capital is being deployed operationally.
Scoring Rationale
Amazon's $17.5 billion covenant-free DDTL facility, set against $200 billion in planned 2026 capex and a simultaneous $10 billion Canadian bond offering, confirms the extraordinary scale of AI infrastructure financing from one of the world's largest cloud providers. The story is directly relevant to ML practitioners tracking compute availability and cloud capacity timelines. Scored just below major strategic announcements since this is a financing filing rather than a product or model release.
Practice with real Retail & eCommerce data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Retail & eCommerce problems


