AI Money Rings: who funds whom, and where to watch
AI demand is real, and it is being monetised: the frontier labs are growing faster than any software company in history, and demand for computing capacity still exceeds supply. Staying invested in this cycle means understanding how it is financed. The build-out is increasingly funded by the companies that sell to it: Nvidia invests in its own customers, Broadcom lends to the lab that leases its chips, and roughly half of the cloud providers’ backlog is owed by just two AI labs. Super-cycles rarely end because valuations look stretched; they end when financing dries up. This map shows where that risk sits, and which companies stand on firmer ground.
How to use: Every arrow follows the money. Tap a company to see its deals and funding stress, or pick a loop to trace it.
Data as of 16 October 2026. Only major transactions between these actors are shown; the map is not an exhaustive record of deals. Funding stress is an indicative assessment of how exposed each company is to capital becoming scarcer or more expensive, based on CDS spreads, credit ratings and recent financing costs. Disc sizes are indicative (market cap or latest valuation, rough ranking). Amounts in US dollars as published; several are caps ("up to") or letters of intent. Rings, from the inside out: chips, AI labs, hyperscalers, neoclouds, financiers and funding vehicles. Not investment advice.