Like a clock, a similar plot unfolds every six months. The &#34;AI bubble theory&#34; always makes its timely appearance, triggering a brief panic in the market, only to be quickly drowned out by a new wave of enthusiasm.
From Goldman Sachs questioning its commercial returns, to China launching highly cost-effective models, to Oracle and OpenAI throwing out a shocking $300 billion &#34;future contract&#34; that stuns the market, the doubts and celebrations surrounding AI alternate in succession.
However, an analysis by Zerohedge reveals that behind this cyclical debate, a deeper structural risk is emerging: the race for AI infrastructure is evolving from a marathon supported by the internal cash flows of tech giants into an &#34;arms race&#34; reliant on external debt.
When a funding gap of up to $1.5 trillion needs to be filled by an already pressured private credit market, one can't help but ask: how far away is that &#34;wolf&#34; that will eventually arrive?
<h2>&#34;AI Bubble Theory&#34; Timely Unfolds</h2>
The first wave of large-scale concerns surged in June 2024. At that time, Goldman Sachs released a report pointing out whether generative AI is a capital black hole of &#34;too much investment, too little return,&#34; a potential pit that may never bring long-term positive returns to investors. This doubt dropped a bombshell in the tech community.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/a04c2046-4fd2-4e62-9f94-16ac49b2c7cb.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/a04c2046-4fd2-4e62-9f94-16ac49b2c7cb.png"/>
However, six months later, after burning another $100 billion to &#34;perfect&#34; the world's most expensive chatbot, a clear profit model still seems to be absent in the United States. In contrast, China launched the widely acclaimed DeepSeek large model, which is not only open-source but also significantly cheaper than its American counterparts, with far less expensive equipment required than the latest NVIDIA super graphics cards.
Meanwhile, reports indicate that companies like Microsoft, Google, and Meta are quietly reducing capital expenditures. These factors combined triggered the next round of sell-off in AI concept stocks, which began in late January and continued until April.
History seems to repeat itself. This inevitably brings to mind the scenes from the first internet bubble, where those once-popular companies ultimately could not escape bankruptcy.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/f764f004-4979-4135-8abf-c994fb80fc1a.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/f764f004-4979-4135-8abf-c994fb80fc1a.png"/>
<h2>&#34;Infinite Money Loophole&#34;: When Financing Shifts from Cash to Debt</h2>
Fast forward to September 2025, the AI bubble is expanding at full speed, pushing the stock market to its highest valuation level since the internet bubble...
However, on September 10, Oracle shattered the tranquility of this celebration with an extremely reckless move. According to a report by The Wall Street Journal at the time, it announced a five-year cloud computing agreement with OpenAI worth a total of $300 billion. This was seen as one of the largest &#34;Vendor Financing&#34; deals in history
More strikingly, Oracle almost simultaneously reminded everyone that it does not actually have enough cash on hand to fund this spending spree, which is expected to last into the 2030s. So, where does the money come from? Borrowing.
JP Morgan analyst Michael Cembalest succinctly described this AI circular economy, referred to by many of his peers as the “infinite money glitch,” in his latest &#34;Market Observations&#34; report.
He explained this phenomenon with a simple circular diagram: AI companies promise to pay large sums to cloud service providers in the future → Cloud service providers borrow against this story to build infrastructure → The infrastructure is then leased back to AI companies.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/91fd5048-1cdf-4924-9129-70effa1d0efd.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/91fd5048-1cdf-4924-9129-70effa1d0efd.png"/>
Cembalest pointed out that since the launch of ChatGPT in November 2022, AI-related stocks have contributed 75% of the S&amp;P 500 index's returns, 80% of its earnings growth, and 90% of its capital expenditure growth. The power consumption of data centers is driving up electricity prices; for example, in the PJM region, 70% of last year's electricity price increase can be attributed to the demand from data centers.
The deal between Oracle and OpenAI is a perfect embodiment of this &#34;glitch.&#34; Doug O'Laughlin from the investment newsletter &#34;Manufactured Knowledge&#34; commented:
<blockquote>
Oracle simply cannot pay for all of this with cash flow; they must issue stock or take on debt to realize their ambitions... A stable oligopoly is breaking apart... What was once a cash flow-funded, disciplined race may now turn into a debt-driven arms race.
</blockquote>
<h2>$1.5 trillion financing gap, can private credit fill it?</h2>
The case of Oracle reveals a deeper issue: the construction costs of AI infrastructure are spiraling out of control and have far exceeded the self-financing capabilities of tech giants. A report from Morgan Stanley paints a shocking picture: it is expected that by 2028, global spending on data center-related expenses will reach approximately $2.9 trillion.
The report notes that while the internal cash flow of large tech companies remains the primary source of funding, after accounting for factors such as shareholder returns, they can self-fund at most about $1.4 trillion. This means the market will face a massive financing gap of up to $1.5 trillion.
Morgan Stanley believes that to bridge this gap, the credit market will play an increasingly important role.
Among all credit channels, private credit is highly anticipated. The bank expects that among the various types of capital to fill the gap, private credit (especially asset financing) will contribute about $800 billion, becoming the most important source of external funding. Consulting firm Bain has also reached a similar conclusion
The future of AI seems to be deeply tied to the money bags of private credit.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/f0719edb-d15c-4624-822f-5dbb32b90439.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/f0719edb-d15c-4624-822f-5dbb32b90439.png"/>
<h2>Private Credit: AI's &#34;Savior&#34; or &#34;Achilles' Heel&#34;?</h2>
However, betting the future of AI on private credit may be a dangerous gamble. Just as the market hopes it will &#34;blood transfuse&#34; AI, the health of the industry itself has raised red flags.
Market data shows that BXSL, a private credit fund under Blackstone Group, one of the largest private credit management companies in the world, has seen its stock price drop to a new low for 2025, significantly underperforming the S&amp;P 500 index. Another industry giant, Blue Owl, is also in a precarious position. According to Bloomberg, Blue Owl has been deeply involved in financing activities in the AI sector.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/101e1f81-54e5-422a-963f-b0f5a8b6f422.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/101e1f81-54e5-422a-963f-b0f5a8b6f422.png"/>
The plight of these private credit giants goes far beyond merely providing funding for data centers. They have been heavily exposed to the weakest links in the U.S. economy—consumers, particularly low-income groups where the non-performing loan (NPL) rates in the &#34;buy now, pay later&#34; (BNPL) sector have soared.
As the Financial Review stated, the private equity industry is &#34;sitting on a $5 trillion survival fear.&#34; If this industry, seen as the financial backbone for AI, falls into trouble, where will the promised $800 billion for AI come from?
<h2>&#34;Bubble&#34; within a Bubble: When No One Talks About the Bubble Anymore</h2>
As structural financial risks become increasingly prominent, public discussions about the AI bubble are cooling down. Deutsche Bank analyst Adrian Cox pointed out that global search volumes for &#34;AI bubble&#34; have dropped by 85% from their peak in August 2025. In other words, the &#34;bubble of discussing the AI bubble&#34; has itself burst.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/fa1a339c-f985-4feb-a9e2-66d16dc24323.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/fa1a339c-f985-4feb-a9e2-66d16dc24323.png"/>
But this does not mean the alarm has been lifted. History shows that the evolution of asset bubbles is not a linear process. In the five years leading up to the bursting of the dot-com bubble in 2000, the Nasdaq index experienced seven pullbacks of over 10%.
More importantly, in November 1998, when investment manager Michael Murphy warned that &#34;this is a serious bubble,&#34; the Nasdaq index was below 2000 points, and it continued to double over the next 16 months, breaking through 5000 points before ultimately crashing.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/0676b605-6729-489f-bad0-8666c84c10a3.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/0676b605-6729-489f-bad0-8666c84c10a3.png"/>
<h2>When Will the Wolf Come?</h2>
After shouting &#34;the wolf is coming&#34; every six months, the market seems to be feeling fatigued. Oracle's massive transaction reveals a dangerous signal of shifting from &#34;practicality&#34; to &#34;borrowing&#34; behind the AI boom, while the anticipated financial backers—private credit—are themselves mired in a quagmire.
This carnival driven by debt and dreams appears even more fragile under the hard constraints of infrastructure like power networks.
Perhaps, when no one talks about bubbles anymore, the wolf has truly quietly arrived at the door. Or, as the old market adage goes: The time the market is irrational always lasts longer than your time to bankruptcy.
So, are we in the largest bubble in history? When will it burst?
The honest answer is: no one knows. Just as NVIDIA's stock price hits a new high and its market value skyrockets to an astonishing $4.5 trillion, the market still buys into the narrative of AI.

GOOGL

NVDA

GOOG

OpenAI

MSFT

ORCL

Every six months, the "bubble theory" arrives as scheduled, triggering a wave of panic and adjustment in the market, only to continue its advance in fervor. But how much longer can this feast last? Dangerous signals seem to have emerged—AI capital expenditures are shifting from "self-financing" to "debt-driven." Morgan Stanley predicts that global data center spending will reach as high as $2.9 trillion by 2028, with approximately $1.5 trillion facing a funding gap that needs to be filled by debt

- The article discusses the recurring "AI bubble" phenomenon, highlighting the cyclical nature of market fears and excitement surrounding AI investments.  
- It emphasizes a structural risk emerging from the shift in AI infrastructure funding from internal cash flows of tech giants to reliance on external debt, with a projected $1.5 trillion funding gap.  
- The potential over-reliance on private credit for AI financing raises concerns, especially as the health of the private credit market itself shows signs of distress.

Wallstreetcn

Alphabet

NVIDIA

Alphabet - C

Microsoft

Oracle

Every six months, there is a wave of "AI bubble theory." When will the "wolf really come"?