ReportGem ReportGem 中文

GLOBAL RESEARCH ARCHIVE

Rating the AI Build-Out: The Datacenter Debt Tailwind for SPGI & MCO

Published: 2026-06-08Institution: EVERCORE ISIPages: 13Original language: 英语Evidence page: 1

Research evidence excerpt

Rating the AI Build-Out: The Datacenter Debt Tailwind for SPGI & MCO

Financials | Business Services

June 08, 2026

David Motemaden, CFA Rating the AI Build-Out: The Datacenter

212-497-0832

David.Motemaden@evercoreisi.com Debt Tailwind for SPGI & MCO

Thomas Gallagher, CFA Bottom Line: Despite recent volatility, we believe that datacenter

212-446-9439 capex will increase as hyperscalers, neoclouds, and datacenter

Thomas.Gallagher@EvercoreISI.com developers build out compute capacity to support increasing adoption

Peter Knudsen of genAI. While genAI fears have weighed on SPGI and MCO as

212-708-8497 investors focus on disruption to non-Ratings businesses, we think the

Peter.Knudsen@evercoreisi.com opportunity on the Ratings side for both to participate in the financing

Nicolas Lu of the AI infrastructure boom is underappreciated.

212-497-0812

Nicolas.Lu@evercoreisi.com We believe that datacenter investment will increase by 50% in ’26 to

$700-800b and by 30% towards $1tn in ’27, with an increasing share

being externally funded with debt that gets rated by SPGI and MCO.

Specifically, we estimate that ~55% of the datacenter capex will be

funded with debt that gets rated by SPGI and MCO at a blended

6.5bps yield, resulting in a 2pt tailwind to Ratings revenue growth

in ’26 & ’27. We see this pushing both SPGI and MCO to the high end

of their Ratings revenue growth guidance ranges for ’26 and

representing upside for ’27, with SPGI in particular being conservative

around hyperscaler debt financing. While positive for the stocks, this

is not a game changer in our base case but we think the risk is

skewed to the upside as a higher debt funded mix, higher revenue

yield or greater AI infrastructure spend could drive a 4-5pts

Ratings revenue tailwind.

The English excerpt is extracted automatically from the cited source page and may contain layout or recognition errors. It is never batch translated.

Open report viewer