JP Morgan: Not just the US stock market, the $5 trillion AI data center boom will also sweep through the US bond market.

According to an analysis by JPMorgan Chase, the actions of artificial intelligence giants in building data centers on a large scale will require an investment-grade bond funding of approximately $1.5 trillion over the next five years, as well as significant financial support from various other sectors of the market. According to a team of strategists led by Tarek Hamid, the question is not "which market will provide funds for the artificial intelligence boom?", but rather "how to design financing schemes to cover all capital markets?" They stated that leveraged financing is expected to provide around $150 billion in funding over the next five years. These strategies, which are complementary, are insufficient to meet the demand, even with support from the investment-grade bond and high-yield bond markets, as well as an annual funding of up to $40 billion from data center securitization. The report estimates that private credit and government funding can make up the remaining $1.4 trillion funding gap. Analysts at the bank wrote in a report released on Monday that this figure is at least $5 trillion and could even reach $7 trillion, which will directly drive the growth of the bond and syndicated loan markets. Analysts expect that next year, there will be $300 billion in high-grade bonds used for building artificial intelligence data centers. This may account for nearly one-fifth of the total issuance volume of the market. A report by Barclays PLC Sponsored ADR estimates that the total issuance volume of this market will increase to $1.6 trillion. The demand for data centers - a demand that analysts say is limited only by physical factors such as computing resources, real estate, and energy - has seen exponential growth in recent months, completely dispelling concerns about the formation of a bubble by some market observers. Last month, Meta (META.US) issued $30 billion in bonds, setting a record for the largest order volume in the high-grade bond market, and last week investors prepared to pay $18 billion to Oracle Corporation (ORCL.US) to fund the construction of a data center campus. Although analysts estimate that the amount of funding needed for expansion in the next five years is staggering, they warn that the road ahead will not be simply "up and to the right." Their biggest concern is that history might repeat itself, similar to the bubble that formed around telecommunications and fiber optic networks in the late 1990s to early 2000s, when oversupply of capacity led to a large number of defaults and a sharp decline in many overvalued assets. In recent weeks, warning signs have become increasingly clear that investors' excessive optimism about data centers may have reached an irrational level. In a recent survey, over half of industry executives expressed concerns about the industry's future challenges, and others on Wall Street also expressed concerns about the complex private debt tools that large-scale enterprises are using to remove artificial intelligence research funding from their balance sheets. Analysts said, "Even if everything goes smoothly, there will still be some big winners, and given the scale of the investment and the competitive nature of certain parts of the artificial intelligence ecosystem (where winners take all), there may be an equal number of losers." According to the bank's strategists, the majority of funding over the next five years will come from these large enterprises themselves. Currently, these companies have net operating revenues of $700 billion per year, with $500 billion being used for capital expenditures.