Amazon reenters the bond market seeking at least $25 billion to support its AI expansion.

      Amazon has returned to the debt markets with a bond sale of at least $25 billion, marking its largest offering this year and highlighting the extent to which the company is willing to borrow to stay competitive in the artificial intelligence sector. The sale comprises eight tranches, with maturities ranging from 2029 to 2066.

      The funds are designated for the major cash-consuming aspects of the industry: data centers, custom silicon, and the infrastructure required for AI. This follows the trend set by Amazon’s record bond issuance in Canadian dollars earlier this year, part of a broader borrowing spree that has exceeded $70 billion since the beginning of 2025 across dollar, euro, and Swiss franc transactions.

      A significant portion of this expenditure is funneled through Amazon Web Services, where the company is quickly expanding its capacity for clients training and operating large models. An increasing amount is also being invested in its Trainium chips, which Amazon promotes as a cost-effective alternative to purchasing Nvidia hardware in large quantities.

      Investor interest was present, although it diminished as the terms became firmer. Orders peaked at approximately $62 billion before the banks managing the sale narrowed the spread, resulting in a final book of about $41 billion, or roughly 1.6 times the size of the bond sale.

      Amazon informed the underwriters that it does not intend to issue additional debt this year, a detail likely intended to reassure a market that is wary of the sector accumulating more leverage. This type of guidance is crucial for investors trying to assess the future volume of debt offerings.

      The rationale for the borrowing is clear on the balance sheet. Amazon has projected around $200 billion in capital expenditures for 2026, which is the highest figure among hyperscale companies, and this has already reduced its free cash flow significantly compared to a year ago.

      The company is not alone in seeking debt to bridge financial gaps. The four largest tech firms in the U.S. are collectively projecting over $650 billion in AI capital expenditures this year, a number that increasingly exceeds their substantial operating cash flows.

      Not everyone perceives the demand as limitless. Analysts have observed that interest in this bond sale appeared subdued compared to previous offerings, suggesting that bond investors are becoming more selective as AI-related debt becomes increasingly available in the market.

      This cautious sentiment is evident in other areas of financing as well. Transactions like Cipher’s junk bonds, intended to finance an Amazon data center in Texas, are testing investors' willingness to support the booming yet riskier segments.

      Amazon structured the bond sale as a combination of fixed and floating-rate notes, allowing for flexibility over a time span that reaches four decades. The longest tranche, maturing in 2066, resembles the ultra-long bonds typically issued by the most creditworthy companies.

      Amazon can still secure these favorable terms because it remains one of the highest-rated corporate issuers, which keeps its funding costs low even as totals increase. The disparity between what Amazon pays and what its smaller competitors pay is becoming a narrative about who can afford to continue expanding.

      The bond sale occurs in a market already saturated with AI-related debt from companies like Meta, Oracle, and various data center developers. Each new issuance raises the question of how much additional debt investors are willing to accept before they begin to demand higher yields.

      For the moment, this bond sale provides Amazon with a financial buffer as it enters a period of significant expenditure with no signs of slowing. The pressing question now for the entire sector is whether the market will remain receptive during the next round of borrowing.