Hyperscalers Alphabet, Amazon, Meta, Microsoft, and Oracle are testing the capacity of global bond markets. In 2025, these issuers were responsible for the year's largest U.S. investment grade bond deals. In September, Oracle sold $18 billion of bonds. In October, Meta sold $30 billion, the largest non-M&A-related bond deal ever. In November, Alphabet followed with $17.5 billion and Amazon with $15 billion, respectively.
A borrowing spree
The pace of hyperscaler issuance is accelerating in 2026 as combined capital expenditure plans for the five largest hyperscalers continue to grow, aggregating to an estimated $750 billion for the full year. On February 2nd, Oracle (Baa2/BBB) priced $25 billion across eight senior unsecured tranches.
On February 10th, Alphabet (Aa2/AA+) coordinated issuance across different currencies and investor bases including $20 billion across seven tranches, £5.5 billion across five tranches including a £1.0 billion 100-year bond, and a CHF 3.1 billion Swiss-franc sale.
Alphabet was not done. On May 5th the company priced its first Canadian issue — also the first Canadian deal by a hyperscaler and the largest-ever Canadian corporate bond issue. The four-part offering of 5-, 7-, 10- and 30-year maturities raised C$8.5 billion. The transaction surpassed Coastal GasLink's C$7.15 billion multi-tranche issue in June 2024 and dwarfed the next-largest Maple bond, Apple's C$2.5 billion single-tranche offering completed in August 2017.
On the same day, Alphabet also priced a €9 billion, six-part Euro offering, bringing combined proceeds raised to $19 billion equivalent without tapping the USD market. The week after the Canadian and Euro financings, Alphabet raised $3.6 billion equivalent in a seven-part Yen-denominated offering. The company's $60 billion, four-month, multi-currency borrowing spree ranks as one of the greatest corporate borrowing binges ever. So far, the ability to issue in multiple denominations, across varied markets, has minimized pressure on spreads in any individual currency or market.
The Alphabet playbook
Alphabet's financings provide a textbook example of global market capacity arbitrage. Tap multiple currencies to diversify buyers and avoid saturating a single market. The ability to issue across multiple currencies, set records for issue size in the sterling and Canadian corporate markets, and place a century bond reflects the substantial market depth that can be surfaced when supply is strategically segmented.
We offer several observations. Liquidity was available and markets held. Across issuers, order books were oversubscribed. The deals priced and cleared at spreads consistent with strong interest. Alphabet's USD tranches priced at spreads to Treasuries of +27bps to +95bps, while lower-rated Oracle required wider spreads of +95bps to +195bps on fixed-rate tranches. The ability for these issuers to repeatedly place very long-duration risk further indicates market depth for issuers whose quality is perceived to be high.
If these transactions were the sum total of issuance from the sector, it would be an impressive volume of funding. But these deals are just a harbinger of the coming wave of supply. Reuters highlights dealer forecasts that U.S. corporate bond issuance could reach historically high levels (Barclays is projecting $2.46 trillion in 2026 and net issuance of $945 billion), with AI hyperscaler issuance a key driver. Reuters also notes the five most active hyperscalers issued $121 billion in aggregate during 2025, far above their $28 billion annual average in the 2020-to-2024 period.
“It is a question of when, not if, other hyperscalers access the Maple and other non-USD markets.”
The coming wave of supply
A concern around market vulnerability is the concentration and cadence of issuers coming to market, not "global market size." Should additional hyperscaler financings come to market in the near term (e.g., multiple $30 to $60+ billion deals within weeks) and/or extend duration materially (40–100 years), the marginal clearing level could shift. Wider new-issue concessions, heavier use of CDS hedges, and more credit spread volatility could result. Both Reuters and the Financial Times have reported investor unease around AI capex payoff timelines, with increased hedging and modest spread widening linked to AI-related bond supply.
In addition to the sheer volume of potential supply, one must consider the structural considerations that affect market capacity. It is not only how much, but what kind of bonds are issued. First, credit quality matters. Second, currency and market segmentation can be effective tools in unlocking additional demand — a capacity-management technique that reduces reliance on any single marginal buyer pool. Finally, issuers need to consider dealer balance sheet and liquidity conditions; even where investor demand exists, dealer syndicates intermediate allocations and facilitate secondary trading.
How much can markets absorb?
The key question is: how much capacity remains at current spread levels before issuers need to start paying up with wider spreads to access liquidity? Global fixed income markets are enormous. The ICE fixed income index tracks more than $100 trillion in government and corporate bonds across 43 currencies. In that context, AI hyperscaler issuance does not seem like a showstopper, and the recent successful financings serve as evidence of ample liquidity in the current market. As to when the market's appetite for this issuance ends — that is not obvious, but we will likely see several additional attempts to test what the market can digest in the coming months.