La inteligencia artificial, una máquina de emitir deuda

As the rise of debt emissions from hyper-scalers disrupts credit and interest rate markets, a radical transformation is occurring within the corporate debt market. Recently, concerns were raised about poor economic data in the United States, and questions were posed about potential market reactions. Interestingly, Barclays noted that global yields have continued to rise despite gloomy economic outlooks in the US.

Possible explanations for this anomaly include the U.S. federal government's significant budget deficit, the emergence of price-sensitive investors demanding adequate compensation, and the new actors attracted to the markets - the hyper-scalers.

These giants are increasingly financing their AI systems through debt markets rather than relying solely on free cash flow. Goldman Sachs reported that USD-denominated corporate bond issuances had already surpassed $1.5 trillion in 2026, with the year projected to surpass the previous record set during the pandemic. Analysts anticipate this figure could be conservative.

The average deal size is $1.7 billion, the largest since the crisis. In 2025, there were 20 transactions of $10 billion or more, a significant increase from the 12 such operations in 2022-2024. While most large transactions are associated with mergers and acquisitions, this trend is different, with two-thirds of the debt issuance this year and last year coming from the technology sector.

The shift in financing methods is causing markets to take unconventional paths. One example is the median weighted term to maturity in USD investment-grade bonds, which is 10.7 years. However, for large tech companies, 20% of issuances have a median term exceeding 14 years, with the five largest tech companies even higher at 16.5 years.

This implies that investors must reconsider their sectoral concentration to avoid the fate of index-linked funds often mirroring active technology investment funds. Moreover, covering these attractive tech bond issuances presents a new coverage challenge, or an alternative to long-term Treasury bonds, for instance.

Lotfi Karoui of Pimco suggests that this is just one of the technical factors influencing the market. Comparing Amazon and Alphabet (Google's parent company) across the euro and dollar markets, he notes that, theoretically, the underlying fundamentals should be practically identical regardless of currency. However, a gap in the performance of performance differentials has emerged, which is difficult to attribute solely to company-specific risks.

Instead, it points to signs of demand fatigue in the USD investment-grade bond market compared to its euro counterpart. Another factor to consider is the construction of indices, as the dominance of a few hyper-scalers in massive bond issuances makes the entire index highly sensitive to a few names, similar to the stock market.

Torsten Sløk of Apollo anticipates that more than $1 billion in financing may migrate towards private placements, including infrastructure loans, asset-backed credit lines, equipment financing, and project-based funding, often accompanied by guarantees, contractual support, and structural protections not available in unsecured public bonds.

He wonders whether this period will be seen as a long-awaited growth phase in an obsolete market, characterized by indices and coverage practices, or a regrettable displacement of debt markets by large technology companies. In any case, the debt market is undergoing transformation before our eyes.