Amazon secures $17.5 billion loan to fund AI expansion

Amazon secured a $17.5 billion delayed-draw term loan, a sharp reminder that the AI infrastructure race is now big enough to pull in even the deepest corporate balance sheets. The facility gives Amazon flexibility to borrow when it needs the cash, while locking in financing for a capital program that is growing more expensive by the quarter.

The senior unsecured credit facility was arranged with Citibank as administrative agent and includes BofA Securities, JPMorgan Chase, HSBC and Wells Fargo among the lenders. Amazon filed the agreement on June 8, and the commitments expire on September 30 unless the company borrows the full amount before then. Any money drawn under the facility carries a three-year maturity from the date it is taken down, and Amazon said the proceeds are for general corporate purposes.

The structure matters as much as the size. A delayed-draw loan lets Amazon secure funding without immediately taking on the full balance, preserving optionality as it decides how quickly to spend on the next phase of cloud and AI expansion. That is useful in a market where companies must commit early to land, power, servers, networking gear and specialized chips long before the revenue from those investments becomes visible.

The borrowing also fits a larger shift across Big Tech. Amazon is not alone in leaning on debt to support AI spending: Alphabet and Meta have both signaled that outlays will remain heavy, and combined AI spending by major technology companies is now expected to top $700 billion this year, up from about $600 billion previously. Meta filed for its largest bond offering ever, up to $30 billion, while Alphabet disclosed plans to sell Japanese yen-denominated bonds for the first time.

Amazon has been layering on financing quickly. Earlier in the same week, the company filed for a five-part debt offering in Canada for as much as C$14 billion, following what was described as the biggest Canadian corporate bond sale on record. Together, the moves show a company making sure it has enough liquidity to keep pace if capital spending accelerates further.

For investors, the signal is clear: AI infrastructure is no longer being treated like a software expense that can be covered comfortably out of operating cash flow. It is becoming a long-cycle industrial investment, with the largest technology companies turning to debt markets to fund the race. Amazon’s latest loan shows just how costly that competition has become.