Schroders analysts report AI-driven capex spurring US hyperscaler bond issuance; active alpha from valuation shifts.

How to play the AI narrative in the public credit market

AI is reshaping credit markets but selectivity matters.

Artificial intelligence (AI) is transforming the global economy and creating fresh dispersion in credit markets. As active multi-credit investors, it is not simply about owning the theme but identifying where AI-driven change is being mispriced and positioning portfolios accordingly.

Why AI risks look different in credit

The market narrative around hyperscalers has shifted from enthusiasm about AI-driven growth to scrutiny of AI-driven capital expenditure (capex). While it is true that the economics of generative AI are still uncertain, we feel that this uncertainty is more acute for equity investors who must assess whether the scale of investment will generate adequate returns, protect margins and justify elevated valuations.

For bondholders, the AI opportunity is fundamentally different from equities. Equity investors need AI investment to translate into stronger growth, margins and valuations. Credit investors have a more defensive hurdle: assessing whether issuers can sustain cash flow, manage leverage and continue servicing debt. On that basis, many of these companies remain among the highest quality borrowers in the market and can offer a more income-focused way to access the evolving AI theme.

Valuations are the key to opportunity

Being able to reassess a portfolio’s exposure and adapt as the AI theme develops is key to capturing these opportunities. Until recently, hyperscalers offered limited appeal for bond investors. As the chart below shows, they were generally a relatively poor source of income with credit spreads tight relative to the overall US investment grade bond market – something that’s even clearer when we exclude Oracle’s significant idiosyncratic effect and materially wider spreads.

US hyperscaler credit spreads versus overall US investment grade market

Source: Schroders, Bloomberg ICE, 30 June 2026 (hyperscalers Amazon, Google, Meta, Microsoft, Oracle). Any reference to regions

securities is for illustrative purposes only and not a recommendation to buy or sell any financial instruments or adopt a specific investment strategy. Past performance is not a guide to future performance and may not be repeated.

That has changed recently. The surge in AI-related capex has driven a sharp rise in bond issuance and contributed to notable hyperscaler credit underperformance. While activity remains largely US-led, issuers are increasingly tapping multiple currencies to fund ambitious capex plans, with 2026 issuance now tracking towards $250 billion (illustrated below: includes Apple and Nvidia). The opportunity lies in being able to respond to these valuation shifts through active positioning.

Yearly global issuance volumes, $billion

Source, Refinitiv workspace. Totals include non-USD corporate bonds. 2026 to 10 August 2026. Any reference to regions

securities is for illustrative purposes only and not a recommendation to buy or sell any financial instruments or adopt a specific investment strategy.

It is remarkable when you consider how much hyperscaler issuer spreads have widened versus the broader US investment grade index. This is particularly stark when you consider their stronger average credit ratings of around AA

AA- (excluding Oracle) compared to the wider US investment grade credit market which averages A- and is heavily weighted to BBB issuers.

In some cases, spreads appear to price in multiple downgrades despite the underlying businesses remaining highly resilient. Heavy issuance has clearly weighed on performance, but the market may be over-discounting AI capex risk.

This is where active management matters as we can distinguish genuine credit deterioration from spread weakness driven by supply, sentiment and equity-market concerns. Such disconnects between valuation and fundamentals can create particularly attractive opportunities to generate alpha.

The active advantage: managing concentration risk

In fact, active credit management is particularly well suited to the AI theme because it allows us to choose where, how much and at what maturity to take exposure, rather than relying on broad index weights and rising concentration risk.

Taking hyperscalers in isolation, these issuers currently represent roughly 4% of the USD investment grade corporate bond market’s market value. That seems small, but their risk contribution is more meaningful than that number represents. Their tendency to issue longer-dated bonds provides greater risk on a DTS 1 basis (illustrated in the chart below) and makes their contribution more impactful. Active management gives us the flexibility to manage these exposures deliberately, avoid overconcentration in the largest borrowers and adjust positioning as supply conditions change.

Average spread duration: US hyperscalers versus US IG excl. hyperscalers

Source: Schroders, ICE BoA US corporate index August 2026. Any reference to regions

securities is for illustrative purposes only and not a recommendation to buy or sell any financial instruments or adopt a specific investment strategy.

Opportunities across the full AI credit ecosystem

The opportunity set is not limited to hyperscalers. Data centre financing, for example, offers another way to access the AI theme, given its critical role in supporting generative AI and cloud computing. Selected issuers can provide attractive premiums, long leases, high-quality counterparties and robust structural protections, allowing us to participate in AI infrastructure growth without relying solely on the debt of the largest technology issuers.

Active management allows us to separate the structural AI theme from the specific credit risk of each issuer and bond. Rather than owning exposure simply because it is in the index, we can assess whether spreads adequately compensate for leverage, duration, supply risk, free cash flow pressure and the credibility of future returns on investment.

That selectivity matters because AI exposure is not uniform. Hyperscalers, semiconductor-linked issuers, data centre owners and wider AI adopters may share the same broad sector labels but their credit drivers, funding needs and sensitivity to AI capex can be very different.

What this means for investors

Within the AI theme, equities may offer greater upside if investment drives a step-change in long-term earnings, but they are also more exposed to margin pressure and shifting market expectations. Credit offers a different profile: more limited upside but attractive income, stronger downside resilience, and exposure to issuers with substantial capacity to service debt.

However, the real opportunity is found in a flexible and selective approach to global credit management, adding exposure where valuations have become attractive, avoiding areas where risks are not properly rewarded and using the broader AI theme (data centres, GPUs, infrastructure and AI adopters) at times when this offers a clearer route to risk-adjusted income.

AI is not the only force influencing credit markets, but it is likely to remain a significant one - and one of many ways we are seeking to generate meaningful alpha for clients.

1 DTS: Duration Times Spread – a measure of a bond’s sensitivity to relative changes in credit spreads.

A version of this article appeared in Investment Week on 25 August 2026

For accredited investors in Singapore only. No part of this material may be reproduced in any manner without the prior written permission of Schroder & Co. (Asia) Limited. Schroders Wealth Management is a trading name of Schroder & Co. (Asia) Limited. Issued by Schroder & Co. (Asia) Limited, 138 Market Street, #23-02, CapitaGreen, Singapore 048946. Registered No: 200719040Z. Schroder & Co. (Asia) Limited is licensed and regulated by the Monetary Authority of Singapore.

Fund Manager, Global Fixed Income

The trillion-dollar question: how much value must AI create to pay for its build-out?

Four reasons why the AI boom is a real estate story

Spotlight on technology: why have AI-related stocks been so volatile?