Author: Bao Yilong
Source: Hard AI
JP Morgan issued a warning that the $5 trillion feast of AI will &#34;squeeze&#34; every credit market.
On November 10, a heavyweight report on the financing needs of AI data centers was released by the JP Morgan strategist team led by Tarek Hamid. The report emphasizes that the construction boom of AI data centers in the next five years will require at least $5 trillion, and it could even reach as high as $7 trillion.
This massive funding will drive the bond and syndicated loan markets to accelerate growth once again. The enormous financing demand means that no financing market can &#34;digest&#34; it alone.
The report predicts that over the next five years, the investment-grade bond market will need to provide about $1.5 trillion, the leveraged finance market about $150 billion, and data center asset securitization can only bear a maximum of $30 billion to $40 billion per year. Even so, there remains a huge gap of $1.4 trillion that needs to be filled by private credit and even government funds.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/0ab6f0fb-86ff-4411-b7fa-87c095a1ba58.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/0ab6f0fb-86ff-4411-b7fa-87c095a1ba58.png"/> (Funding sources for AI/data centers)
<h2>A Capital Feast</h2>
The research report points out that the global construction of AI and data centers will be an &#34;extraordinary and sustained capital market event.&#34;
The foundational forecast of the report shows that between 2026 and 2030 alone, the world will need to add 122 gigawatts of data center infrastructure capacity. A more optimistic forecast based on semiconductor orders suggests that the growth scale in the next three years could reach 144 gigawatts.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/0070897f-505b-47e1-b93e-624b483af201.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/0070897f-505b-47e1-b93e-624b483af201.png"/> (The relationship between data center installed capacity and annual capital expenditure)
However, this feast faces &#34;hard constraints&#34; from the physical world, with electricity being the biggest bottleneck.
The delivery cycle for new orders of gas turbines has been extended to 3-4 years, while the construction cycle for nuclear power plants can take over ten years.
Balancing the new electricity demand while managing residential electricity prices will become a politically sensitive economic issue in the U.S. Although these physical limitations slow down the construction speed, they also mean that without these constraints, the funding demand would be even larger.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/4f1275ec-4327-4092-a7bb-a67fe5d35e9b.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/4f1275ec-4327-4092-a7bb-a67fe5d35e9b.png"/> (Average electricity prices for U.S. residents)
<h2>Trillion-Dollar Funding Sources</h2>
Such a massive capital expenditure is far beyond what a single market can bear. JP Morgan believes that the future will be a funding pyramid constructed by different capital markets, with each layer playing an indispensable role.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/9f4f5255-90b6-4b34-978e-178e394faeac.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/9f4f5255-90b6-4b34-978e-178e394faeac.png"/> (Capital expenditure growth trend of hyperscale cloud service providers)
First, the cornerstone is the cash flow of tech giants.
Large tech companies generate over $700 billion in operating cash flow each year, with about $500 billion reinvested in capital expenditures. JPMorgan Chase assumes that approximately $300 billion of this cash flow will be directly used for investments in AI and data centers each year.
Secondly, the main force is the high-grade bond market, which will bear most of the financing &#34;responsibility.&#34;
JPMorgan Chase expects that the high-grade bond market can absorb about $300 billion in AI-related bonds over the next year, and this figure will accumulate to $1.5 trillion over the next five years.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/c65eb1d8-b1cc-4d58-b240-afc086ae383a.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/c65eb1d8-b1cc-4d58-b240-afc086ae383a.png"/> (Most hyperscale cloud computing companies are rapidly increasing their debt levels)
Currently, AI and data center-related industries account for 14.5% of the JULI index, surpassing the U.S. banking industry. By 2030, this proportion may exceed 20%.
The JULI index is used by JPMorgan Chase to measure the performance of the investment-grade dollar-denominated corporate bond market. In simple terms, it is an important indicator that tracks the overall market for high-quality corporate bonds in the U.S.
Next is the supporting force of leveraged finance and the securitization market.
The leveraged finance market (high-yield bonds and leveraged loans) has the capacity to provide about $150 billion in funding over the next five years. However, the report also issued a historical warning: in the 1990s, the telecom industry became the largest sector in the high-yield bond market before facing a collapse; the expansion of the energy sector from 2010 to 2015 also ended in disappointment.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/9292d851-9812-4652-a558-827ae128317d.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/9292d851-9812-4652-a558-827ae128317d.png"/> (Currently, the technology sector accounts for 16% of the leveraged loan index and 7% of the high-yield bond index.)
The securitization market is the &#34;natural home&#34; for data center financing, expected to absorb $30 billion to $40 billion in risk capital each year. However, its main function currently is to provide construction financing rather than permanent financing, which somewhat limits its role.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/b0c141ce-8ddc-4480-a1f3-bc28664d87a5.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/b0c141ce-8ddc-4480-a1f3-bc28664d87a5.png"/> (The scale of securitized products related to data centers has grown by 83% year-on-year, currently accounting for 5% of the total issuance of ABS and CMBS)
Finally, attention needs to be paid to the key &#34;fillers,&#34; private credit and alternative capital.
After utilizing all public markets, JPMorgan Chase estimates that there is still a significant funding gap of about $1.4 trillion. This gap will mainly be filled by private credit and alternative capital
The private credit market has a funding pool of approximately $466 billion, and its structural flexibility is extremely high, allowing for tailored solutions for complex projects.
Recently, Meta completed a $27.3 billion private financing through a tool called &#34;Beignet Investor LLC,&#34; which is a typical innovative case that cleverly moves construction funds and long-term lease obligations off the balance sheet.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/43649e32-5141-46bd-ae7b-e9be55c255f4.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/43649e32-5141-46bd-ae7b-e9be55c255f4.png"/> (Organizational structure of Beignet Investor LLC)
<h2>Historical Echoes in Frenzy: Warnings from the Telecom Bubble</h2>
The report also compares the current AI craze with the telecom and fiber optic network construction bubble around the year 2000:
<blockquote>
Similarities:
<ul>
<li>Back then, the market firmly believed that internet data would &#34;double every 100 days,&#34; leading to massive investments in fiber optic networks. Many financially troubled internet bubble companies and over-leveraged telecom builders participated.</li>
<li>Today, the exponential growth expectation for AI computing power demand is driving a similar investment frenzy.</li>
</ul>
Key Differences:
<ul>
<li>The fundamental reason for the collapse of the telecom bubble was that &#34;the revenue curve failed to keep up with the investment curve.&#34; The slow growth in demand for high-speed networks from end users and businesses could not support the returns on massive investments, leading to bankruptcies of companies like Global Crossing and the collapse of the industry.</li>
<li>Today's tech giants like Amazon, Google, Meta, and Microsoft have extremely strong free cash flow and asset-heavy balance sheets, and their financial strength is far beyond that of companies during the internet bubble period.</li>
</ul>
</blockquote>
Nevertheless, the report believes that this history remains a powerful warning: no matter how promising the technological prospects, if they cannot be converted into tangible revenue, large-scale capital investments may ultimately leave nothing but a mess.
<h2>Winner Takes All and Inevitable Losers: The Ultimate Risk of Investment</h2>
At the end of the report, JPMorgan Chase pointed out two core risks that will ultimately determine the outcome of this capital feast.
The first is monetization risk. JPMorgan estimates that to achieve a 10% annual return on AI investments over the next five years, approximately $650 billion in new revenue needs to be generated each year.
This figure is equivalent to 0.58% of global GDP, or about $34.72 per month for each iPhone user worldwide. Although the primary payers will be businesses and governments, the enormity of this figure reveals the daunting path to monetization.
The second is disruptive technology risk. Analysts believe that an investment frenzy, if based on a specific technological path, is extremely sensitive to sudden changes in technological efficiency.
The report cites the &#34;DeepSeek moment&#34; as an example, where a startup claimed to achieve top model performance at a very low cost, causing market panic over concerns that existing expensive GPU investments would quickly become obsolete, turning into &#34;dark fiber&#34; (fiber that has been laid but not used)
JP Morgan's report clearly indicates that the wave of AI infrastructure construction is irreversible, and it will inject unprecedented vitality into the capital markets.
However, in this over $5 trillion gamble, not all participants will emerge as winners. The &#34;winner-takes-all&#34; nature of the AI ecosystem means that even in the most optimistic monetization scenarios, some companies will inevitably be eliminated.
For investors, understanding the flow of funds, identifying companies with real moats and monetization capabilities, and maintaining a sense of awe towards historical bubbles will be key to succeeding in this epoch-making game of opportunities and risks

GOOGL

MSFT

GOOG

JPMorgan Chase's strategist team believes that the construction boom of AI data centers in the next five years will require at least $5 trillion, with the investment-grade bond market needing to provide about $1.5 trillion, the leveraged finance market about $150 billion, and data center asset securitization up to $40 billion annually. Even so, there remains a huge gap of $1.4 trillion that needs to be filled by private credit and even government funding

- Morgan Stanley warns that the $5 trillion AI investment boom will strain every credit market.  
- The report highlights a projected funding need of $5 to $7 trillion for AI data centers over the next five years, with significant gaps requiring private credit and government funding.  
- It draws parallels to the telecom bubble, emphasizing the risks of monetization and disruptive technology that could impact the success of AI investments.

Over $5 trillion! JPMorgan Chase: Global AI infrastructure is "unprecedented in scale" and will impact all capital markets