Goldman Sachs Research says hyperscale cloud providers will spend $5.3 trillion on AI and data-center capital expenditures from 2025 through 2030, a rise from an earlier $4.5 trillion estimate, and warns this scale is beginning to strain liquid credit markets and issuer concentration limits. Goldman’s chief credit strategist Amanda Lynam cautioned that as hyperscalers account for a growing share of new corporate borrowing, public debt investors may become less willing or able to absorb ever-larger amounts from the same handful of issuers.
Goldman argues private infrastructure and real estate funds will play an increasingly important role in filling financing needs, noting infrastructure funds raised a record $221 billion in 2025 and held roughly $400 billion of dry powder as of September 2025. The firm projects private infrastructure AUM could exceed $3 trillion by 2030 if current fundraising trends continue, reflecting faster fund closings and stronger investor interest in infrastructure’s income and inflation-protection characteristics.
Morgan Stanley offers complementary analysis, forecasting about $2.9 trillion of global data-center construction and hardware capex through 2028 and identifying roughly a $1.5 trillion external financing gap after accounting for hyperscalers’ internal cash generation. Morgan Stanley breaks that gap into likely channels including private credit (especially asset-based finance), securitization and bond issuance, and sovereign or private equity capital, with private credit expected to be a dominant source.morganstanley
Goldman also warns that capex estimates are rising faster than actual data-center construction, which could shift future bottlenecks from model demand to financing capacity, power supply, and project execution underscoring the importance of diversified financing across markets, structures and currencies. The complexity of data-center financing spanning land, power, networks, buildings and equipment means any market correction could transmit losses through a more fragmented set of investors and instruments than in prior tech buildouts.
Read Article: CISF Proposes NATGRID-Linked Facial Recognition Network for Airports
