Amazon has mandated banks for its inaugural sterling bond issuance, which is expected to launch as early as Wednesday. Meanwhile, global hyperscale technology giants are continuously raising capital to support artificial intelligence infrastructure development.
Zhitong Finance APP has learned that the U.S. e-commerce and cloud computing supergiant $Amazon (AMZN.US)$ has mandated banks to arrange its inaugural pound-denominated bond issuance, with the transaction expected to launch as early as Wednesday. Currently, global hyperscale cloud service providers are continuously raising debt worldwide to fund artificial intelligence infrastructure construction. An unprecedented "AI credit" boom is shifting the financing landscape for AI computing infrastructure of U.S. tech giants from the U.S. dollar market to the global capital pool. Previously, Amazon raised approximately EUR 14.5 billion through an eight-tranche bond issuance in March, and is now further preparing for its debut pound-denominated bond offering.
Giants such as Google and Amazon have been establishing a durable and diversified sovereign currency-based hyperscale market financing structure this year to minimize the negative impacts of AI bubble panic and prevent the depletion of liquidity reserves supporting AI computing infrastructure expenditures.
Alphabet, the parent company of Google, completed its inaugural issuance of Australian dollar-denominated corporate bonds in August. Also in August, the company completed a $25 billion bond issuance in the U.S. dollar market. Furthermore, it issued corporate bonds denominated in Swiss francs, pounds sterling, euros, Canadian dollars, and Japanese yen in 2026. The company recently also raised nearly $85 billion through an additional equity offering.
Alongside peers such as Meta Platforms (parent company of Facebook), Oracle, Amazon, and even Microsoft, these U.S. tech supergiants have already raised hundreds of billions of dollars this year through financing in U.S. dollars and other sovereign currencies to support their artificial intelligence ambitions.
Although massive investments in artificial intelligence by large technology companies are helping to push AI-computing-themed tech stocks back to historic highs, the backdrop of substantial corporate bond issuances has intensified market debate. Higher corporate bond yields and the yield curve for U.S. Treasuries with maturities of 10 years or more have fueled arguments over whether these companies can generate sufficient profits from such vast scales of AI computing infrastructure investment. Moreover, risks associated with a potential burst of the "AI bubble," driven by the interplay of continuously increasing corporate bond issuance volumes, rising yields, and default expectations, are escalating.
Financing Landscape Expands Again! Amazon Knocks on the Door of the Sterling-Denominated Bond Market for the First Time
Alphabet, the parent company of Google, and Amazon are converting their AA-grade credit quality, strong cash flows from search/e-commerce/cloud computing businesses, and robust revenue growth from the sale/leasing of AI computing resources into sources of low-cost long-term global capital. They are diversifying across multiple currencies and investor bases to avoid saturating the single U.S. dollar bond market with massive AI-related supply. While these latest developments in the AI credit market significantly weaken panic surrounding the "AI bubble/unsustainable financing" narrative specific to Alphabet, they do not completely dispel the broader controversy regarding bubbles across the entire AI industry.
According to a person familiar with the matter who requested anonymity due to the non-public nature of the information, the company plans to issue multiple tranches of corporate bonds with maturities ranging from 3 to 19 years. JPMorgan, Barclays, HSBC Holdings, and NatWest Group are mandated to arrange the transaction.
This online retail giant and the world's largest cloud computing provider made its debut in the eurobond market in March, setting a historical record for the volume of euro-denominated corporate bond issuances. Subsequently, the company entered the Swiss franc bond market, issuing a record six tranches of bonds. Overall, Amazon and Alphabet have jointly become the largest bond issuers among hyperscale cloud service providers in 2026. The total amount of securities issued by Amazon, converted into U.S. dollars, exceeds $92 billion, slightly higher than that of Alphabet.
Amazon's total debt also ranks first among hyperscale cloud computing giants. Microsoft, which trails Amazon in cloud computing market share, ranks second globally; however, Amazon's debt scale is nearly double that of Microsoft, which follows closely behind. Amazon plans unprecedented capital expenditures of $220 billion this year, with CEO Andy Jassy stating that the majority will be used to expand AI-related computing infrastructure resources.
Given the unprecedented scale and speed of such bond issuances, financial market investors have shown persistent weakness in the pricing trajectory of tech giants' equities and bonds since the beginning of the year. Demand for debt investments attracted by recent issuance transactions appears to have weakened. Coupled with widening spreads during pricing as long-term U.S. Treasury yields rise, this is driving up the cost of debt financing.
Global capital is flooding into the AI credit market, subjecting the boom in global AI computing power investment to a major test of returns.
Amazon raised €14.5 billion through an eight-tranche bond issuance in March and is now preparing for its inaugural pound sterling bond offering. Research published by the European Central Bank on August 31 indicates that the outstanding euro-denominated bonds issued by hyperscale cloud service providers amount to approximately €40 billion, with large U.S. technology companies accounting for nearly 10% of the new issuance volume of euro-denominated non-financial corporate bonds this year. Cross-currency financing can expand the investor base, alleviate the pressure of concentrated issuances, and match overseas expenditures. It also helps mitigate the risk of an AI bubble bursting through diversification of debt currencies and investor bases. However, the actual cost advantages depend on bond coupons, maturities, and the costs associated with currency hedging and cross-currency swaps.
Alphabet’s practice of multi-currency financing is particularly notable: in February, it issued £5.5 billion in bonds, including a century bond with a principal of £1 billion, a coupon rate of 6.125%, and maturity in 2126. In August, it entered the Australian dollar bond market for the first time, raising A$5.5 billion (approximately $3.89 billion), with subscriptions exceeding A$18 billion, more than 3.27 times the issuance volume. The 20-year bonds carried a coupon rate of 6.9%. From ultra-long-term capital in the UK to local currency bond investors in Australia, the expansion of AI infrastructure is absorbing savings from diverse markets.
Financing instruments are also expanding. In February, Oracle announced a plan to raise $45–50 billion by 2026, split roughly equally between debt and equity financing, to build additional capacity for contracted cloud computing demand. In August, NVIDIA announced partnerships with institutions such as Apollo, BlackRock, and Blackstone to establish an independent computing power financing platform, aiming to gradually mobilize over $500 billion in third-party capital, subject to final agreements. These developments indicate that AI computing infrastructure construction is simultaneously connecting tech companies' balance sheets, public bond markets, and institutional capital, making financing capability a critical determinant of expansion speed.
However, the 10-year U.S. Treasury yield, approaching 5%, continues to impose a rigorous stress test on AI capital expenditure financing. With U.S. federal debt surpassing $40 trillion and the 10-year Treasury yield rising to approximately 4.8%, Ruchir Sharma, Chairman of Rockefeller International, views a "decisive breakthrough above 5%" in yields as a significant signal that could impact the AI boom. He points out that the transmission mechanism lies in increased government borrowing needs and rising fiscal risk premiums, which may push up benchmark financing costs. Consequently, tech giants must offer higher returns to continue attracting the capital required for building AI infrastructure.
He contrasts annual AI application revenue of approximately $200 billion with infrastructure investments exceeding $1 trillion, emphasizing the financing pressure caused by upfront investment and delayed revenue realization. While these two figures measure application revenue and infrastructure scale respectively and cannot be directly subtracted to determine the industry's funding gap, the impact of rising external financing costs is clearly quantifiable: for every $100 billion in new borrowing, a 100-basis-point increase in interest rates adds $1 billion to the annual interest burden. Meanwhile, higher discount rates reduce the present value of future cash flows, placing long-construction-cycle projects that rely on continuous financing under dual pressure on profitability and valuation.
Edited by Deng