Higher borrowing costs to split AI buildout by credit quality, Morgan Stanley says

What happened

Morgan Stanley strategists said rising borrowing costs are unlikely to derail the artificial intelligence infrastructure buildout but could widen the divide between stronger and lower-rated companies, with the four largest hyperscalers expected to raise total capital spending by more than 50% in 2027 from 2026.

Confirmed

Global impact / market context

Higher borrowing costs matter most for companies with less cash and weaker credit, because interest payments take a bigger share of budgets and can force cuts to AI spending. Stronger firms can keep funding projects, so the buildout may continue but concentrate among a few large players.

Analyst inference

The four largest hyperscalers directly or indirectly account for over 75% of expected AI-related capital spending, so their plans drive much of the sector. For them, costs may rise through wider credit spreads, which means investors demand higher interest over safer debt, rather than less borrowing.

Analyst inference

What to watch

1. Morgan Stanley expects the four largest hyperscalers to increase total capital spending by more than 50% in 2027 from 2026, so their spending plans are central to the buildout. Confirmed
2. Watch whether lower-rated hyperscalers, neoclouds, and other AI participants slow projects or borrow less as funding costs rise, which would show the divide widening. Proposed
3. Track credit spreads, which mean the extra interest paid over safer debt, for AI-related borrowers; wider spreads would raise costs even if the strongest companies keep issuing debt. Analyst inference

Evidence

• seekingalpha.com