Tech issues: The AI debt deluge hitting bond markets

In 2025, five of the largest hyperscalers – Amazon, Alphabet, Meta, Microsoft and Oracle – alone issued $121 billion in US corporate bonds versus an average of $28 billion per year between 2020 and 2024¹.

This level of issuance is likely not a one-off, but could represent the beginning of a longer-term trend as AI investment continues to accelerate. Worldwide spending on AI is forecast to come to $2.52 trillion in 2026, a possible 44% increase year-over-year, according to Gartner². This compares to a total of $1.6 trillion spent on AI between 2013 and 2024. Thus far, tech companies have funded their rapid expansion through free cash flow, however, with elevated bond market appetite and relatively cheap funding costs, these companies are tapping into bond markets to fuel the seemingly inexhaustible capex demand, as well as to optimise their capital structure to meet payments to shareholders.

“This level of issuance is likely not a one-off, but could represent the beginning of a longer-term trend as AI investment continues to accelerate.”

By October 2025, the amount of debt tied to AI had ballooned to $1.2 trillion, making it the largest segment in the investment grade market. At 14% of the high-grade market, AI now surpasses US banks as the largest sector in the JP Morgan US Liquid index³. Given this scale, and fears over an AI ‘bubble’, how should bond investors approach this surge in tech issuance?

Quality and quantity

The AI ‘hyperscalers’ have tended to be investment grade (IG) companies, with strong balance sheets and low levels of leverage.

In fact, there is often a question that some of these tech giants might be considered more ‘risk-free’ than the traditional ‘risk-free rate’ of US Treasuries – thanks to their low debt levels, ability to generate significant revenue and integrity to today’s society. This perspective is reflected in the tightness of spreads of some of these companies over government bonds.

This convergence between government and credit markets in risk terms is highlighted by Alphabet’s decision to issue a century bond. 100-year bonds are rare and more often issued by governments than corporates as a century is a long period, introducing significant uncertainty; the market needs to be confident in the longevity of the business model and its ability to adapt over time. For example, Motorola issued a century bond in 1997. At the time of issuance, Motorola was investment grade and a dominant player in technology. But due to tech disruption, its core franchise collapsed and the company was split and restructured. The bonds didn’t default, but the business that investors thought they were lending to no longer exists in its original form. Despite this, Alphabet received close to ten times orders on £1 billion worth of century bonds. While the sterling-denominated bond is set to have a higher credit rating than the UK government itself⁴, it could be an example of AI exuberance tipping into bond markets.

However, despite the fundamental quality of these companies, the sheer volume of current tech issuance has created somewhat of a dislocation – the bonds of these high quality issuers can be bought at a spread pick up over less fundamentally strong credits. As a result, the market is currently very happy to take advantage of this surge in issuance.

“The bonds of these high quality issuers can be bought at a spread pick up over less fundamentally strong credits.”

Even following this capex cycle, the sheer scale of these companies limits the leverage impact – even after billions of issuance, some of these giants will geared 0.4 – 0.7 times, compared to an average leverage of just under three times for the US IG market⁵. Tapping debt markets reflects capital structure optimisation and shareholder‑return commitments, not a shortfall in internal resources. Furthermore, these hyperscalers know they cannot flood the market with debt. For example, Oracle scaled back its debt requirement by using equity on balance sheet concerns.

What are the risks?

A rerating of the broader IG market is possible as investors may reallocate from other issuers into new hyperscaler deals, particularly if they offer attractive new-issue concessions. However, while the absolute volume of issuance is large, it is not necessarily at a level that would create systemic pressure or crowd out the wider market. Any valuation impact would thus be more flow driven (contingent on investor demand) than fundamentally driven. However, the potential contagion from a tech valuation unwind into wider credit markets is a risk worth monitoring.

“There are likely to be bumps in the road, and for an active manager, this can present buying opportunities.”

Furthermore, the current capex announcements are coming in a very liquid market. However, we cannot assume that there will always be such liquidity. There are likely to be bumps in the road, and for an active manager, this could present buying opportunities.

Within tech specifically, a notable area to monitor is the risk of a correction in AI-related valuations and a growth outlook that may not fully materialise. As bond investors, there is an implicit assumption is that issuers will have sufficient cashflow to pay the coupons and to repay the principal to refinance at maturity; we are most interested in their balance sheet. We begin to worry when companies engage in behaviours that disguise genuine liabilities on the books – we see this occur in most credit cycles, indicating a hot part of the credit cycle. Given this, it is important to take a selective approach, taking credit risk when sufficiently compensated, and when we aren’t, increasing exposure to the healthiest credits and in more senior parts of the capital structure.

¹ Reuters, ‘AI hyperscalers will drive higher US corporate bond supply in 2026’, (reuters.com), January 2026.
² Gartner, ‘Gartner says worldwide AI spending will total $2.5 trillion in 2026’, (gartner.com), January 2026.
³ Bloomberg, ‘At $1.2 trillion, more high-grade debt now tied to AI than banks’, (Bloomberg.com), October 2025.
⁴ Bloomberg, ‘Alphabet 100-year bond gets most demand in sterling sale’, (Bloomberg.com), February 2026.
⁵ JP Morgan, December 2025.