Mapping global data center geography to US‑reported cloud revenues and profits

Determine how the global geographic distribution of data centers operated by US‑headquartered cloud infrastructure providers (Amazon Web Services, Microsoft Azure, Google Cloud Platform, IBM Cloud, Oracle Cloud, and CoreWeave) maps into the revenues and profits recognized by US‑domiciled subsidiaries, and ascertain the extent to which cross‑border intra‑group transfer pricing influences those reported figures, to accurately assess domestic value added in national accounts.

Background

The paper explains that hyperscale cloud providers headquartered in the United States build and operate data centers worldwide, with roughly half of global capacity estimated to be located in the US. However, the precise country distribution and, critically, how revenues and profits are recorded within multinational corporate structures are not publicly reported with sufficient clarity.

Understanding how geography, corporate organization, and transfer pricing shape reported revenues and profits is crucial for determining how much value added from globally distributed data center operations appears as domestic output in the United States versus being attributed to other jurisdictions. This has direct implications for interpreting national accounts and measuring the AI sector’s macroeconomic footprint.