Amazon.com is moving toward a groundbreaking financing structure to secure the computational resources necessary for the AI race without directly weighing down its balance sheet. According to reports, the cloud and e-commerce giant is considering an $8 billion deal to acquire Nvidia's top-tier processors using external investment capital.
The SPV Model and Asset-Light Strategy
The proposed deal involves transferring thousands of advanced Grace Blackwell chips—currently installed in more than 12 US data centers—into a Special Purpose Vehicle (SPV). This financial engineering allows Amazon to adopt an "asset-light" capital structure.
Under a sale-leaseback model, the SPV will own the hardware while Amazon leases it back for operational use. The investment vehicle is expected to issue debt and offer up to 10% of its equity to third parties. If it achieves an investment-grade rating, it is anticipated to attract long-term capital from pension funds and insurance groups.
Massive CAPEX and Hardware Lifecycles
Amazon's move reflects the immense capital requirements faced by global Hyperscalers. The company has already announced that its total capital expenditures (CAPEX) will reach a staggering $220 billion this year, with the vast majority directed toward generative AI infrastructure and GPUs.
A key driver for this shift is the rapid pace of technological evolution. Processors in advanced data centers typically require replacement every five years, making traditional outright purchases increasingly expensive and financially rigid for corporations maintaining massive scale.
