The global debt issuance related to AI has already reached nearly $236 billion by May 31, 2026, Morgan Stanley estimated.

Morgan Stanley sees global debt issuance for AI companies to top $500 billion in 2026 on the back of massive capex plans of top hyperscalers like Amazon, Google and Meta.

The investment bank forecasts global debt issuance could double to $570 billion as AI companies tap alternative sources for funding for their growth plans, according to a Reuters report.

A number of big-bang funding plans have been announced recently for the AI buildout. Google parent Alphabet recently announced an $85 billion fundraising for its expansion of AI capabilities, including data centres and compute facilities. It also issued a rare 100-year bond in February for its AI funding requirement.

The global debt issuance related to AI has already reached nearly $236 billion by May 31, 2026, Morgan Stanley estimated, reported Reuters. This is a fourfold increase from last year.

The booming demand for frontier AI models, along with the focus to ramp up agentic AI, has led to AI companies like Anthropic and OpenAI announcing major spending plans. Both have confidentially filed for their public market debut. While Anthropic is valued at $965 billion, OpenAI is valued at around $852 billion.

The massive compute infrastructure needed to train these AI models has led to more capacity additions by hyperscalers and plans to spend even more capital to ramp up capacity. Morgan Stanley estimates hyperscaler capex to top $1 trillion in 2027. It said that these hyperscalers are broadening their base by issuing more non-USD issuances.

Amazon recently issued C$14 billion ($10.04 billion) of Canadian dollar-denominated notes, Reuters reported, citing a final pricing term sheet filed with the SEC. The company that runs the cloud services AWS raised 14.5 billion euros ($16.88 billion) from an eight-part deal, the largest ever in the euro corporate bond market.

Hyperscalers are diversifying their funding as they plan to spend trillions of dollars on creating data centre capacity to match the compute demand in future.

Published on June 10, 2026