JPMorgan: The construction cost of AI is projected to reach $7 trillion, and the AI investment frenzy may repeat the dot-com bubble.

JP Morgan's latest report indicates that many AI hyperscalers are ramping up their data center expansions. This infrastructure boom is expected to involve over $5 trillion in funding over the next five years, potentially reaching as high as $7 trillion. Analysts estimate that investment-grade corporate bonds alone will need to raise $1.5 trillion, a massive amount that could almost require “moving the entire global bond market out” to sustain it. Despite the influx of capital, JP Morgan warns that the AI investment frenzy could repeat the mistakes of the 2000s dot-com bubble.

A wave of debt is emerging, and AI financing exceeds the burden of a single market.

JPMorgan pointed out that the funding demand for AI infrastructure is enormous, far exceeding the capacity of any single market, stating:

“The question is not which market should fund AI, but how to design a financing structure that can simultaneously open the doors to all capital markets.”

According to estimates, investment-grade corporate bonds will contribute about $1.5 trillion over the next five years, the leveraged finance market can provide an additional $150 billion, and data centers raise about $40 billion annually through securitization.

Nevertheless, there is still a funding gap of 1.4 trillion dollars, which may need to rely on private credit and government investment to fill.

Total costs soar past 7 trillion dollars, and the bond market will expand again.

J.P. Morgan estimates that the total cost of AI infrastructure will be at least $5 trillion and could reach as high as $7 trillion, becoming a major driving force for the expansion of the global bond and syndicated loan markets.

Analysts predict that investments in investment-grade corporate bonds related to AI data centers will reach $300 billion next year, accounting for about one-fifth of the overall issuance in the high-rated bond market. Barclays (Barclays) adds that the size of the high-rated bond market will grow to $1.6 trillion by 2026, indicating that capital is pouring into the AI infrastructure sector.

The demand explosion has become the norm, and large tech giants are competing to issue bonds.

JPMorgan pointed out that the demand for AI data centers has entered a “explosive” growth phase, with the only limitations currently being physical conditions such as land, electricity, and computing resources. Recent large bond issuance cases show that the market heat is astonishing:

Meta issued $30 billion in corporate bonds last month, setting the record for the largest order in investment-grade bond history.

Oracle has also attracted investors to rush in, willing to invest an additional $18 billion to support data center construction.

Overall, investors' confidence and enthusiasm for AI infrastructure continue to rise, with no signs of cooling in the market in the short term.

The market is a bit overheated, and AI investments may repeat the internet bubble.

Although the prospects for AI development are enormous, JPMorgan also warns that future developments will not be a straight path upward. They are concerned that this wave of enthusiasm may resemble the internet bubble of the early 2000s, when overbuilding led to an oversupply, ultimately triggering a wave of defaults and a collapse in valuations.

In recent weeks, there have been constant warnings of an overheated market:

More than half of the data center operators admit that they are worried about potential financial pressure and industry turmoil in the future.

Wall Street is also uneasy about AI giants' operations to “offload assets from the balance sheet” using complex private debt structures.

JPMorgan warns:

“Even if the progress is smooth, this competition will ultimately only produce a few astonishing winners, accompanied by a similarly tragic batch of losers.”

Super company digs deep into its pockets, investing $500 billion each year for expansion.

JPMorgan pointed out that the primary source of funding in the next five years will still come from these “AI super companies” themselves. These tech giants currently have an annual operating profit of about 700 billion USD, of which 500 billion USD is reinvested in capital expenditures for expanding data centers and AI facilities.

In other words, AI construction not only siphons off liquidity from the global bond market but also allows these giant companies to sustain the entire AI construction almost solely with their own funds, becoming the largest financial backers in the market.

(The Economic Reality Behind the AI Craze: Two Data Centers in California Have Been Idle for Years Due to “Power Shortages”)

This article Morgan Stanley: AI construction costs are projected to reach 7 trillion USD, and the AI investment craze may repeat the internet bubble, first appearing in Chain News ABMedia.