Arthur Hayes: Both AI Debt Outcomes End in Money Printing

Hayes is framing AI infrastructure debt not as a private credit story but as a sovereign backstop question, which matters because it shifts the risk from hyperscaler balance sheets toward public institutions. The claim is structural rather than predictive: if compute demand is underwritten by state capacity, the buildout becomes a fiscal and monetary project, not a market one. The open question is whether policymakers ever acknowledge that role explicitly, or whether any backstop arrives only after a demand shock forces it.