In the current bond market, a significant portion of issuance—about a quarter—is driven by artificial intelligence (AI) investments, primarily to fund data centers. Miriam Wheeler from Goldman Sachs highlights that this trend is not just a passing phase but represents a generational opportunity centered on hard assets and infrastructure buildout. The demand for power and data center capacity is expected to double in the U.S. over the next few years, with hyper scalers planning to spend $600 billion on AI-related capital expenditures this year alone. This surge is influencing various market segments, including investment grade, bank market, high yield, and structured products, with high yield markets particularly dominated by AI-related issuance.
Despite the enthusiasm, there are natural constraints such as land and power availability and long project lead times that act as checks on the pace of development. Wheeler anticipates that as more data becomes available, the leveraged finance market will begin to differentiate between projects based on their track record and ability to deliver on time and budget. This differentiation will be crucial as investors seek to manage risks associated with the uncertain long-term impact of AI, which could either disrupt society profoundly or face regulatory and operational challenges.
The current market frenzy is fueled by strong investor appetite and available capital, but Wheeler notes that project finance fundamentals remain important. Factors such as tenant credit quality, lease length, and project location play a significant role in investment decisions. While the first quarter saw robust M&A activity, much of it was concentrated in sponsor and strategic deals, with a relatively light backlog moving forward. This suggests a positive technical outlook for the market, with expectations that sponsor activity will return, particularly driven by AI infrastructure and other hard asset sectors like healthcare and industrials.
Outside of AI and hard assets, market activity is expected to be muted, especially in software, which remains a challenging sector with a wide bid-ask spread between buyers and sellers. However, sectors like staples and healthcare are showing more promise, and European markets are becoming more active again. Wheeler points out that while software is not toxic, transaction volumes are limited until valuation gaps narrow. Credit spreads remain tight despite inflation and geopolitical tensions, reflecting a bifurcated market where clean credit stories perform well, but more sensitive sectors face downward pressure.
Looking ahead, there is cautious optimism about the economy’s ability to absorb inflationary pressures without triggering broader market disruptions. Wheeler notes that strong earnings in the first quarter and the economy’s resilience have supported this stability. While concerns remain about potential contagion from sensitive sectors, so far, the market has managed to isolate these risks effectively. Overall, the AI-driven infrastructure buildout presents a compelling investment theme that is shaping the bond market’s dynamics and investor focus for the foreseeable future.
