Blab AI
Hyperscalers Push AI Debt into New Currencies as Bond Sales Reach Record Levels
Amazon and Alphabet have issued more than $60 billion of bonds in multiple currencies over the past year, setting new records for corporate debt in euros, yen, Canadian dollars, Swiss francs and sterling. The moves reflect the growing capital‑expenditure needs of the world’s largest cloud and data‑center operators as they expand artificial‑intelligence (AI) infrastructure.
Amazon’s March sale of €14.5 billion – the largest ever in the euro corporate bond market – was part of an eight‑part deal that raised $16.6 billion in total. Alphabet’s recent issuances broke records in several markets: a yen bond, a Canadian‑dollar bond, a Swiss‑franc bond and a sterling bond all set new highs for their respective currencies. Alphabet also sold a 100‑year bond, the first tech‑company issue of that maturity since 1997. According to LSEG data, these transactions reshaped the global bond market and expanded the pool of investors beyond the United States.
The debt issuances come as hyperscalers’ capital‑expenditure budgets climb. BNP Paribas estimates that hyperscalers will spend about $725 billion on AI‑related infrastructure this year, roughly double the level seen in mid‑2025. Analysts say that operating cash flow is growing slower than the pace of spending, creating a need for external funding.
Bankers are also experimenting with new financing structures for AI startups and data‑center operators. In early July, Stingray Compute – a unit of Cipher Digital – issued an $810 million note that was nine times oversubscribed. The note was backed by a pre‑arranged lease to Amazon, a structure inspired by construction loans. Morgan Stanley’s Cody Gunsch said the first deals of this type began last year and about 15 have since been sold to high‑yield investors.
Despite the surge in debt supply, investors are questioning whether the market can absorb further issuance. Morgan Stanley’s Teddy Hodgson warned that AI debt could push investment‑grade issuance above $2 trillion in 2026, the first time the figure would exceed the total for the year. Investment‑grade deals from hyperscalers have already surpassed their full‑year 2025 total and are on pace to reach BNP Paribas’ $250 billion forecast for 2026.
Crossmark Global Investments’ Victoria Fernandez noted that the bonds are high‑quality and that liquidity remains strong. She added that repeated issuances could raise concerns about market saturation. Barclays’ Scott Schulte said that AI‑related debt in the United States accounts for about 15 % of total investment‑grade issuance, but that the share is still low relative to the broader market.
Manulife Investment Management’s Jeff Given said hyperscalers are still ramping up long‑term AI projects, keeping the funding pipeline open. Alphabet’s CFO Anat Ashkenazi confirmed the company has accumulated $100 billion in outstanding debt across six major currencies, while CEO Sundar Pichai has said the company funds investments through a mix of cash flow, debt and equity.
The trend of multi‑currency bond sales reflects a strategy to tap a wider investor base and avoid saturation in the U.S. dollar market. It also signals that large technology firms are willing to diversify their debt profiles to support AI‑driven growth.
In summary, hyperscalers are issuing record‑size bonds in a variety of currencies to finance AI infrastructure, while banks are testing lease‑backed structures for data‑center operators. Demand for AI debt remains strong, with analysts projecting that investment‑grade issuance could exceed $2 trillion in 2026. No signs of market saturation appear at present, but the continued growth of AI spending and the need for external financing keep the debt market under close scrutiny.
Amazon’s March sale of €14.5 billion – the largest ever in the euro corporate bond market – was part of an eight‑part deal that raised $16.6 billion in total. Alphabet’s recent issuances broke records in several markets: a yen bond, a Canadian‑dollar bond, a Swiss‑franc bond and a sterling bond all set new highs for their respective currencies. Alphabet also sold a 100‑year bond, the first tech‑company issue of that maturity since 1997. According to LSEG data, these transactions reshaped the global bond market and expanded the pool of investors beyond the United States.
The debt issuances come as hyperscalers’ capital‑expenditure budgets climb. BNP Paribas estimates that hyperscalers will spend about $725 billion on AI‑related infrastructure this year, roughly double the level seen in mid‑2025. Analysts say that operating cash flow is growing slower than the pace of spending, creating a need for external funding.
Bankers are also experimenting with new financing structures for AI startups and data‑center operators. In early July, Stingray Compute – a unit of Cipher Digital – issued an $810 million note that was nine times oversubscribed. The note was backed by a pre‑arranged lease to Amazon, a structure inspired by construction loans. Morgan Stanley’s Cody Gunsch said the first deals of this type began last year and about 15 have since been sold to high‑yield investors.
Despite the surge in debt supply, investors are questioning whether the market can absorb further issuance. Morgan Stanley’s Teddy Hodgson warned that AI debt could push investment‑grade issuance above $2 trillion in 2026, the first time the figure would exceed the total for the year. Investment‑grade deals from hyperscalers have already surpassed their full‑year 2025 total and are on pace to reach BNP Paribas’ $250 billion forecast for 2026.
Crossmark Global Investments’ Victoria Fernandez noted that the bonds are high‑quality and that liquidity remains strong. She added that repeated issuances could raise concerns about market saturation. Barclays’ Scott Schulte said that AI‑related debt in the United States accounts for about 15 % of total investment‑grade issuance, but that the share is still low relative to the broader market.
Manulife Investment Management’s Jeff Given said hyperscalers are still ramping up long‑term AI projects, keeping the funding pipeline open. Alphabet’s CFO Anat Ashkenazi confirmed the company has accumulated $100 billion in outstanding debt across six major currencies, while CEO Sundar Pichai has said the company funds investments through a mix of cash flow, debt and equity.
The trend of multi‑currency bond sales reflects a strategy to tap a wider investor base and avoid saturation in the U.S. dollar market. It also signals that large technology firms are willing to diversify their debt profiles to support AI‑driven growth.
In summary, hyperscalers are issuing record‑size bonds in a variety of currencies to finance AI infrastructure, while banks are testing lease‑backed structures for data‑center operators. Demand for AI debt remains strong, with analysts projecting that investment‑grade issuance could exceed $2 trillion in 2026. No signs of market saturation appear at present, but the continued growth of AI spending and the need for external financing keep the debt market under close scrutiny.
