Bond navigators: AI issues dominate market

The world is obsessed with AI. For the past two years it’s been hard to get away from it. Equity markets have risen on a wave of AI optimism. Now credit markets are trying to absorb a wave of issuance as hyperscalers and related areas look to raise capital to fund the necessary investment.

Cloud-based service providers – generally known as hyperscalers – have been busy raising capital to pay for ambitious investment plans. The numbers being mentioned seem fantastic - $1.5 trillion over the next five years[1] has been mooted, with overall numbers needed to fund overall AI infrastructure several times that number.

We saw the start of this last year, with Oracle ($18 billion), Meta ($30 billion), Alphabet ($17 billion) and Amazon ($15 billion) all coming to market in the second half of 2025 [2]. There has been no let-up in the first weeks of 2026.

A boost to sterling credit markets

Oracle announced a $25 billion issue in February – with the company keen to highlight that this would be its only deal this year. The issue was split into eight parts – all US dollar, spanning maturities from three years to 40 years, and was more than 4x subscribed.

One point of interest is how these deals are being structured. Hyperscalers have recognised that while the US dollar market is deeper than others, the amounts required necessitate a broader approach. Put another way, they are tapping as many pools of capital as possible to meet their funding needs. In February, we’ve had Google parent company Alphabet raise money across US dollar, sterling and Swiss franc markets. In the sterling market, the company raised more than £5 billion across five tenors ranging from three years to 100 years – the latter more than 10x subscribed.

The use of the sterling market is noteworthy: it showed that there are plenty of investors happy to lend to this sector, and that borrowers can extend maturity here perhaps more easily than is the case in the US dollar market.

Cheap bonds… becoming cheaper?

Thus far, we’ve seen bonds come to market at attractive levels – for instance the recent Alphabet  sterling 30 year bond deal  priced at around 120bps over equivalent gilts, when we might usually see an AA rated firm to issue at 60-70 bps. It’s interesting that post-sale, the bonds have generally been trading slightly softer. While I think fundamentals are sound on these issues, there are two technical factors that are probably weighing on these in the short term, namely that with the bonds relatively cheap, I think that some investors were expecting a tightening post-issue, and that the market is looking ahead and recognising that we are still near the beginning of this process and that there will be a lot more supply in the coming months.

What’s coming next?

The market is already looking at the next potential mega-deal, but we expect to see smaller AI-related infrastructure deals as well. We’ve already seen examples of bond issues to fund data centres. We’ve participated in a number of these deals: many data centres being built already have long-term contracts to provide capacity to hyperscalers, and the bonds they issue are often shorter term and secured, which can provide an attractive risk / return profile. As ever, there is more than one way to play a new theme for active investors.

[1] https://www.mellon.com/insights/insights-articles/record-breaking-ai-related-debt-issuance-in-2025.html
[2] https://www.reuters.com/business/finance/ai-hyperscalers-will-drive-higher-us-corporate-bond-supply-2026-analysts-say-2026-01-15/

This is a financial promotion and is not investment advice. Past performance is not a guide to future performance. The value of investments and any income from them may go down as well as up and is not guaranteed. Investors may not get back the amount invested. Portfolio characteristics and holdings are subject to change without notice. The views expressed are those of the author at the date of publication unless otherwise indicated, which are subject to change, and is not investment advice.