Bank of America raises its ultra-large-scale computing corporate debt forecast by 25% to $175 billion

MaticHoleFiller · 2026-03-28T09:32:43+00:00

Bank of America raises its 2026 forecast for ultra-large-scale computing enterprise investment-grade bond issuance to $175 billion, indicating that the AI financing cycle may accelerate. Amazon's $54 billion bond issuance has sparked strong demand, and it is expected that Meta, Microsoft, and Alphabet will issue $30 billion, $20 billion, and $15 billion in bonds respectively. Investment in ultra-large-scale computing enterprises is projected to grow by 70%, reaching $600 billion, and may continue beyond 2026.

MaticHoleFiller

2026-03-28 09:32:43

Abstract generation in progress

Bank of America raised its forecast for the issuance of investment-grade bonds by hyperscale computing companies for 2026 by 25%, to $175 billion, indicating that the AI financing cycle may expand faster than previously expected. Before this revision, Amazon completed a $54 billion multi-currency bond transaction related to its investment in OpenAI, which was not included in earlier estimates. Analysts currently expect roughly an additional $65 billion of issuance for the remainder of this year; if a large portion of the supply has already entered the market, this level could support credit spreads.

So far, hyperscale computing companies have issued about $110 billion of investment-grade bonds in 2026, accounting for about 63% of Bank of America’s updated forecast. Issuance channels remain active, and Meta Platforms Inc. is expected to issue about $30 billion, Microsoft to issue $20 billion, and Alphabet Inc. to issue $15 billion. About one quarter of the expected total issuance is likely to be denominated in foreign currencies, reflecting a broader global funding strategy. Amazon’s recent deal also attracted strong demand; it was reported that U.S. orders reached $126 billion, underscoring that even as the scale of supply continues to grow, investor interest remains undiminished.

The broader backdrop points to a sharp expansion in AI-related capital expenditure. Investment by hyperscale computing companies in data centers is expected to rise by about 70% year over year, to more than $600 billion, raising questions about how long this pace can be sustained. As cash-flow levels after capital expenditure catch up to those before capital expenditure following shareholder returns, further increases in spending guidance could translate directly into higher borrowing demand. At the same time, signals from Oracle about future spending needs may indicate that issuance trends could continue beyond 2026, keeping the credit market tightly connected to the trajectory of AI infrastructure investment.

A wealth of information and precise analysis—available in the Sina Finance app

Byline: Zhang Jun SF065

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

1 Likes