The ongoing establishment of Artificial Intelligence (AI) infrastructure is no longer exclusively shaping the equity market but is also having a significant impact on fixed-income markets. Major technology corporations, known as “hyperscalers”, are increasingly funding their expansive investments through debt. Tilmann Galler, Capital Markets Strategist at J.P. Morgan Asset Management, states that AI development is not only transforming working and living environments but is also fundamentally reshaping capital markets.
The resulting record issuance from US technology companies is having a lasting impact on the structure of relevant credit indices. This has consequences for passive bond investors, as the risk profile changes. In this context, active security selection is gaining importance. The necessity of financing AI infrastructure has pushed issuance activity in the US bond market to historic levels. In the current year 2026, issuance by US hyperscalers already amounts to approximately 200 billion US dollars.
This amount represents 15 per cent of all gross new issuance in the US investment-grade segment and exceeds three times the average volume of the past five years. Economist Tilmann Galler noted that technology companies had already projected a further increase in investment volumes for 2027 in their recent quarterly reports. This development is manifesting across markets: currently, 18.7 per cent of all new high-yield issuance is attributable to the technology sector. Although the sector's weighting in the overall high-yield index is still comparatively low at 8.6 per cent and below the 10 per cent weighting in the investment-grade index, Galler interprets this as a structural consequence of the AI investment cycle, whose duration is linked to continued infrastructure spending.
Concentration Risks in the Bond Market
The issuance boom introduces a phenomenon to the bond market that investors primarily know from equity markets: concentration risk. Unlike equity indices, which are weighted by market capitalisation, bond indices are weighted based on the volume of outstanding debt. Capital Markets Strategist Galler explains that passive investments in this environment are increasingly made in companies with high debt burdens, regardless of the actual soundness of the business model. This implies that bond indices ultimately 'reward' higher indebtedness, which represents a significant concentration risk for diversified portfolios.
A negative market event, such as disappointing revenue growth or a reduction in AI investments, would simultaneously affect the largest positions in bond and equity indices. Since index-linked funds must mirror the weighting of their benchmarks, increased risk premiums (spreads) could trigger reallocations that would exacerbate a downward movement. Tilmann Galler also points out that the fundamentals of hyperscalers appear robust at first glance, and metrics such as the debt-to-EBITDA ratio seem unproblematic. However, a more detailed analysis reveals risks, as investments in data centres were often structured via leasing agreements. These leasing liabilities are not accounted for during the construction phase but only upon commencement of use.
Risk Reduction Strategies
Considering these 'hidden debts', the creditworthiness situation becomes relative. This correlates with the increase in premiums for credit default swaps (CDS) of hyperscalers in recent months, as Galler explained. Investors are thus faced with the question of whether the achievable return still offers adequate compensation for the increasing risk in credit indices and whether an increasing concentration in the technology sector within the overall portfolio is sustainable. To mitigate concentration risks, Galler recommends supplementing with other sectors in the investment-grade market, as well as bonds with shorter and medium maturities, to reduce sensitivity to interest rates and spreads.
- —Supplementing the portfolio with bonds from other sectors in the investment-grade market.
- —Including bonds with shorter and medium maturities to reduce interest rate and spread sensitivity.
- —Utilising securitised credit, whose returns derive from diversified pools of collateral.
- —Implementing active bond strategies to specifically reduce concentration risks in the technology sector.
Furthermore, securitised credit represents a qualitative option, as its returns are generated from various pools of collateral and are thus independent of companies' AI expenditures. Active bond strategies offer the opportunity to specifically reduce concentration risks in the technology sector because, unlike passive approaches, they do not necessarily have to follow benchmark weighting with increasing indebtedness but can individually weigh the risk-reward ratio, Galler concluded.
