Amazon is reportedly exploring a financing deal that would move about $8 billion worth of Nvidia chips off its balance sheet without taking those chips out of service. The plan, first reported by the Financial Times and summarized by Bloomberg and Yahoo Finance, would let Amazon keep using the hardware while outside investors help finance it.
According to Bloomberg's account of the FT report, Amazon is exploring a structure that would place thousands of Nvidia Grace Blackwell chips used across more than a dozen U.S. data centers into a special-purpose vehicle. Yahoo Finance's writeup of the FT report said Amazon would then lease the chips back, effectively keeping the compute capacity while changing who finances the assets.
That would fit a more asset-light model for AI infrastructure. Bloomberg said the vehicle could raise money by issuing debt and selling up to a 10% equity stake to investors, while Yahoo Finance similarly reported that outside investors could fund the structure through debt issuance, with Amazon also planning to offer an equity stake of up to 10%.
Yahoo Finance, citing the FT, said the move is aimed at strengthening Amazon's balance sheet and comes amid a broader push by hyperscalers to find less balance sheet-heavy ways to fund giant data center expansions.
Amazon's spending plans help explain why this kind of structure is under discussion. Yahoo Finance reported that the company has said it expects to spend more than $200 billion in capital expenditures this year, with much of that going to Amazon Web Services for chips and new data centers. In that context, alternative ways to finance AI hardware would reflect how large the infrastructure bill has become.
The key qualifier is that this is still a reported exploration, not an announced transaction. The underlying reporting describes a deal Amazon is seeking or exploring, with talks held in recent weeks, rather than a completed sale or finalized financing arrangement.
If it happens, the deal would be another example of Big Tech trying to keep feeding its AI appetite without letting every chip purchase sit directly on the balance sheet.
