Morgan Stanley sees AI financing as defining summer 2026

Morgan Stanley analysts predict AI financing will dominate summer 2026, rather than new technological advancements. The four leading hyperscalers - Microsoft, Alphabet, Amazon, and Meta - are projected to boost total capital expenditures by 57% in 2027 compared to 2026. This surge in investment is based on the expectation that these expenditures can yield a return on invested capital of at least 25%.

However, the disparity between capital deployment and revenue generation is straining near-term cash flow. Morgan Stanley's 2027 free cash flow estimates for these hyperscalers have decreased, widening the financing gap and indicating a need for increased AI-related credit issuance. Credit spreads for hyperscalers have widened, particularly for higher-quality unsecured bonds, with issuance volumes surging.

While the major hyperscalers, with average ratings around AA, have substantial financing requirements and flexible ratings, lower-rated issuers like Oracle and data center developers have less financial flexibility and a lower tolerance for higher funding costs. The upcoming phase of AI financing is anticipated to focus on compute equipment, including servers and chips, as well as energy assets.

Notably, Nvidia's introduction of a compute infrastructure financing platform and a $35 billion chip financing transaction backed by Broadcom are indicative of this trend. Morgan Stanley anticipates high-quality issuers to provide more financial support, such as backstops, credit support arrangements, and residual value guarantees, to facilitate private capital in financing larger pools of AI infrastructure assets.