Amazon's $80B AI Hardware Loan Facility Moves Capex Into Asset-Backed Credit

SEATTLE / LONDON (BYTEVYTE) - Amazon has disclosed a secured loan facility of up to $80 billion that lets it borrow against AI hardware it buys. The AI hardware loan facility can be drawn from early 2027 through the third quarter of 2028, and Amazon disclosed it alongside its quarterly results. The structure matters because it moves part of the funding for AI infrastructure away from the corporate bonds and operating cash the company has relied on, and toward lenders secured on the equipment itself.

Amazon decides when to draw on the facility. Each draw carries a variable rate and a five-year maturity, and Amazon will grant lenders a security interest in certain purchased assets. The filing does not name the specific assets that would serve as collateral.

The size of the commitments explains the timing. Total principal contractual commitments reached $883.6 billion at the end of September, up 36% from the prior quarter, and Amazon's 2026 capital expenditure forecast stands at $220 billion. Long-term debt climbed to $173.1 billion from $65.6 billion at the close of 2025, about 2.6 times the year-start level.

How the AI hardware loan facility shifts risk

Over the 12 months ending September 30, free cash flow came in at negative $42.4 billion. Amazon is putting more into capital projects than its operations produce, and the facility offers another means of bridging that shortfall. So far, Amazon has paid for data center construction and chip purchases using cash from its cloud and retail businesses and through its largest-ever corporate bond sales. The new line introduces a source that does not rely on immediate cash or conventional debt issuance.

The arrangement is also described as supplier financing, which ties the credit directly to hardware purchases. That linkage changes who carries the downside. Lenders hold collateral on servers and chips, assets whose value depends on continued demand and on the pace at which newer chip generations replace them. If utilization weakens or hardware loses value faster than the five-year maturities, the lenders' recovery depends on assets that the filing does not identify. Amazon, by contrast, keeps discretion over the timing of each draw.

Trade-offs and open questions

The facility offers Amazon flexibility. It can draw in tranches as capacity comes online, and it avoids a single large bond issue in a period when debt has already grown sharply. The costs are rate exposure from variable pricing and a claim on specific hardware. The filing and the coverage so far do not disclose the spread over benchmark rates, the fees, or the amount Amazon expects to draw in any given period.

The facility also funds assets whose footprint extends beyond the balance sheet. The sources describe the money as targeting data center construction and AI chips, but they do not report the power, water, or emissions profile of the sites that this capital will build. Many AI data centers depend on on-site gas turbines, which are loud for people living nearby. Gas-fired power carries high CO2 emissions, and cooling these facilities consumes large volumes of water. When financing becomes easier to obtain, the local costs of the build-out can grow faster than the public disclosures about them, and residents near new sites have little visibility into those costs in the materials provided.

Verdict

The facility is a financing tool that lets Amazon keep expanding AI capacity while its free cash flow remains negative and its debt has more than doubled since the end of 2025. The risk moves toward lenders who hold collateral on fast-depreciating hardware and toward the suppliers whose sales depend on the draws. Decision-makers who extend credit to, or sell hardware into, large cloud providers should ask which assets secure these lines and how collateral values are tested. The next checkpoints are the first draws from Q1 2027, the assets Amazon eventually pledges, and whether free cash flow turns positive against the $220 billion capital expenditure forecast.

While we strive for accuracy and rigor, ByteVyte can make mistakes. Readers are advised to verify all information independently and conduct due diligence before making decisions based on this reporting. We accept no liability for errors, omissions, or decisions made in reliance on this content.

Sources

AMAZON COM INC - Form ARS - FY2023

Related Articles

• Amazon's $8 Billion Nvidia Chip Sale-Leaseback Turns GPUs Into an Asset Class
• Amazon Projects $200 Billion AI Infrastructure Spend as AWS AI Revenue Hits $15 Billion Run Rate
• BIS Flags AI Capex Financing Risk as Debt Replaces Cash Flow

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.