The "burn rate" of AI infrastructure may reach $5.5 trillion. JP Morgan: The bond market has the capacity to absorb the wave of new debt, and tech giants can still leverage further.

As technology giants launch a wave of bond issuance to build AI data centers, the market begins to worry about whether the U.S. investment-grade bond market can absorb the growing supply of debt. However, Stephanie Aliaga, a global markets strategist at J.P. Morgan Asset Management, believes that the current leverage levels of the largest cloud computing companies remain relatively low, and the strong demand for AI computing power also supports cash flow in the future. Therefore, the bond market is fully capable of absorbing the new issuances. J.P. Morgan estimates that the six largest cloud service providers could even add approximately $1.5 trillion in debt on top of their current levels without putting significant pressure on their financial status. Currently, bonds from these six major cloud computing companies account for about 5% of the U.S. investment-grade bond index, a figure that has doubled from two years ago. As investments in AI infrastructure continue to expand, the influence of these tech giants in the global bond market is rapidly increasing. In an interview on Tuesday, Aliaga stated, "We believe they will continue to issue bonds." Bonds from the six AI giants have doubled in proportion over two years, with an additional $1.5 trillion available for bond issuance. Building AI data centers requires massive capital investments, prompting tech companies to increasingly turn to the bond market for long-term funding. Currently, the six largest cloud service providers have already captured about 5% of the U.S. investment-grade bond index, doubling from two years ago. The rapid growth in bond supply has led some investors to start worrying about whether tech companies will increasingly exert supply pressure on the investment-grade bond market. However, J.P. Morgan Asset Management believes these companies' balance sheets still have ample room for expansion. Aliaga points out that, compared to the overall investment-grade bond market, the leverage ratios of the six largest cloud service providers are still significantly lower. J.P. Morgan estimates that these companies can easily add approximately $1.5 trillion in debt on top of their existing levels. The bank anticipates that as AI infrastructure investments continue to expand, large tech companies will continue to utilize the bond market for financing. Aliaga stated that debt itself is not a bad thing. For the super-sized cloud providers building data centers expected to be operational for five, ten, or even longer years, debt may actually be a very attractive means of financing. As U.S. Treasury yields rise to their highest levels since 2008, the wave of bond issuance by tech giants raises concerns of "buyer insufficiency." As Aliaga expressed these views, the global bond market is facing significant pressure. Concerns about persistently high U.S. inflation and increased expectations for further interest rate hikes by major central banks have led U.S. government bond yields to reach their highest levels since 2008. At the same time, the momentum in AI infrastructure construction is driving large tech companies to accelerate the issuance of investment-grade bonds. The simultaneous expansion of government financing needs and tech company financing demands has raised investor concerns about whether the bond market has enough buying power to absorb the new supplies. However, Aliaga believes this concern may be exaggerated. She stated, "We believe the market is fully capable of absorbing these new bond issuances. If there is any impact, it will likely make the AI boom more sustainable." In other words, if the bond market can continue to provide large tech companies with long-term, relatively stable financing channels, the construction of AI data centers will not have to rely solely on corporate cash flows, thereby helping to extend the current cycle of AI capital expenditures. Anthropic's computing contracts have exceeded $175 billion, and the scale of investment in AI infrastructure continues to swell. Computing power for AI remains one of the most scarce resources in the global tech industry. For example, the AI company Anthropic has already committed to signing cloud computing contracts worth more than $175 billion, highlighting the enormous demand from AI businesses for computing infrastructure. As large tech companies, AI labs, and cloud computing enterprises compete for GPUs, servers, storage, networking, and data center resources, the capital needs throughout the entire AI industry chain are rapidly expanding. This financing demand is not only pushing tech companies to increase bond issuance, but also beginning to compete with the financing needs of the U.S. government and other sovereign issuers for funds in the global fixed income market. However, Aliaga believes that the AI demand itself is also providing important support for this debt. Currently, the operating cash flow of super-sized cloud providers can roughly cover their capital expenditures. However, what is more noteworthy is that the backlog of customer contracts signed by the three largest super-sized cloud providers is growing at a faster pace than capital expenditures. Aliaga considers this a positive signal, as it indicates that the enormous investment by these companies in AI infrastructure is garnering increasing support from future customer demand, thereby enhancing the likelihood of achieving investment returns on these projects. Global AI infrastructure investment could reach $5.5 trillion by 2030. J.P. Morgan predicts that by 2030, global AI infrastructure investment could reach a staggering $5.5 trillion. Such a vast capital demand means that even the largest tech companies with strong cash generation capabilities will find it challenging to bear the entire investment solely through their operating cash flows. Aliaga stated that the cash flows of super-sized cloud providers can only cover part of the $5.5 trillion investment demand. Therefore, future debt financing and other forms of external capital will play an increasingly important role. In addition to the public bond market, alternative capital sources such as private credit may also become important funding sources for AI infrastructure financing. This means that the AI investment boom may increasingly extend from the tech stock market to the bond and private credit markets in the future. For investors, assessing the AI capital expenditure cycle will no longer only involve judging how much tech companies are willing to invest. They will also need to pay attention to whether the global capital markets can continue to provide sufficient financing for these projects. The AI capital expenditure cycle will eventually slow down, after which profit margins and free cash flow are expected to improve. Aliaga believes that the current massive capital expenditures by super-sized cloud providers will not maintain such rapid growth indefinitely. As AI infrastructure is gradually built, the growth rate of capital expenditures will ultimately slow down. At that time, the profit margins and free cash flow pressures of large tech companies are expected to ease. She remarked that the eventual deceleration of capital expenditures "should provide some relief for profit margins and free cash flows." However, at least in the foreseeable future, the issue of tight AI computing supply is still unlikely to be resolved completely. Thus, the truly critical question in the next stage of the AI investment cycle may not be whether demand exists, but rather who can first overcome the computing supply bottlenecks and when these newfound capacities will truly come online. Aliaga specifically pointed out that memory supply bottlenecks are one of the significant limitations currently facing the expansion of AI infrastructure. For investors, it will be crucial going forward to assess which companies can first overcome such supply constraints for essential components like memory and bring new AI computing power to market.