AI earthquake on the stock market: Why $800 billion burned up in just one week – and hardly anyone noticed?

A stock market earthquake: The November collapse of the AI ​​sector

The artificial intelligence system has collapsed, and no one seems to have noticed—or rather, many have noticed and are now busy counting the wreckage of their investments. In the first week of November 2025, the technology sector experienced a dramatic collapse that not only shattered the euphoria of the previous months but also raised fundamental questions about the economic viability of the entire AI infrastructure boom. The numbers are so monumental they are almost incomprehensible: Eight of the most valuable AI-focused companies lost nearly $800 billion in market capitalization in a single week. The Nasdaq Composite Index fell 3 percent in five trading days, its weakest performance since the tariff turmoil of spring 2025. This was not a correction of an overbought market but the moment investors became aware of an uncomfortable reality: The assumptions on which the entire valuation cascade rests may not be sustainable.

Suitable for:

• “The German Angst” – Is the German innovation culture backward – or is “caution” itself a form of sustainability?

A systematic loss of trust

The market capitalization of Nvidia, the world's most valuable company, shrank by approximately $350 billion in just five trading days. Palantir Technologies, which had surged 374 percent this year, lost more than 10 percent of its value after releasing its quarterly results, even though the figures exceeded expectations. Oracle, Meta, and AMD—all major players in the AI ​​ecosystem—experienced similar declines. This wasn't a selective drop in individual overvalued companies, but rather a systemic loss of confidence in the entire AI infrastructure thesis.

Suitable for:

• The $57 billion miscalculation – NVIDIA of all companies warns: The AI ​​industry has backed the wrong horse

The impossible calculation: Trillion-dollar investments without a business model

The parallels to historical speculative bubbles are too striking to ignore. Back during the dot-com era, companies invested hundreds of billions of dollars in building fiber-optic networks under the oceans, hoping that applications would emerge later. They were fundamentally wrong in their assumptions about the necessary infrastructure. Something similar is happening today, but on an even larger scale. Tech giants Alphabet, Amazon, Meta, and Microsoft spent a combined $112 billion on AI investments in the third quarter alone. Meta CEO Mark Zuckerberg expects costs to reach approximately $600 billion by 2028. OpenAI and Oracle have announced plans to invest $500 billion in a so-called Stargate data center project. Amazon has announced it will spend more than $30 billion on investments in each of the next two quarters. The sheer sum of these expenditures – Bain estimates that annual capital expenditures will reach approximately $500 billion by 2030 – raises the central question: What revenues must be generated to justify these expenditures?

The search for profit: Why the AI ​​revolution isn't making money yet

The answer Bain Capital provides in its analysis is as sobering as it is illuminating. To justify just the investments made in data centers in 2023 and 2024, the industry will need annual revenues of approximately two trillion dollars by 2030. That's many times more than realistic scenarios predict. It's more than Apple, Amazon, Alphabet, Microsoft, Meta, and Nvidia earned combined last year. It's more than five times the entire global software market.

Profitability as a mirage

The profitability of these massive capital investments remains entirely speculative. An analysis conducted by Bain revealed that 95 percent of early corporate AI initiatives have yet to generate a profit. OpenAI itself—the flagship company of the generative AI movement—is generating approximately $13 billion in revenue this year, while paying Oracle an average of $60 billion annually for data center capacity. This means OpenAI would have to increase its revenue sixfold just to meet its Oracle contracts before any talk of profitability even arises. This is not a business model, but a structured calculation that depends on a continuous flow of fresh investment into the system.

Circular finance: How the industry is inflating itself

The structural problem is exacerbated by the circular financing dynamics. Nvidia has pledged to invest $100 billion in OpenAI, expecting OpenAI to use those funds to purchase Nvidia's GPUs for data centers. This is a classic Ponzi scheme, where value is artificially inflated through reciprocal investments between the same players. Meta has secured $29 billion in funding from investors such as Pimco and Blue Owl Capital, not through operating profits, but through promises of future success. Oracle had to sell $18 billion in bonds to finance its data center expansion plans. CoreWeave, a data center company that went public in March, has raised $25 billion since last year through public debt and equity markets to fund its own expansion. Tracing this financing chain reveals not a stable business model, but a fragile structure that relies on the markets' continued willingness to accept debt and buy stock at record valuations.

Mathematically absurd: Palantir, Nvidia and the rating frenzy

The valuation situation is unhealthy. Palantir trades at a price-to-earnings ratio of 313, meaning an investor would have to spend 313 years earning their money back to earn its market capitalization. Nvidia, even with solid profitability compared to other AI companies, operates in a valuation regime that is difficult to justify even under optimistic assumptions. When the company announced in October that Nvidia had no plans to sell its new Blackwell chips in China for the time being and was not in “active discussions with China,” its stock lost $229 billion in market capitalization in a single day. This illustrates the extreme reliance on specific geopolitical and strategic narratives, rather than fundamental business metrics.

The bear is loose: Michael Burry's billion-dollar bet against the hype

The most prominent financial investor, Michael Burry, known for predicting the 2008 housing market collapse, placed massive bets against Nvidia and Palantir through his investment firm, Scion Asset Management. In September 2025, Burry purchased put options on approximately 5 million Palantir shares worth $912 million and on 1 million Nvidia shares worth $187 million. It is quite striking that the disclosure of these positions, made as part of regulatory 13-F filings, coincided immediately with a massive market movement. An investor like Burry, who has proven his ability to identify bubbles before they burst, would not randomly bet his capital against the hottest growth stocks of the year. The fact that the markets did not register these signals earlier speaks to the speculators' psychological state: they simply didn't want to see them.

A fragile foundation: Macroeconomic headwinds for the tech industry

Simultaneously with this valuation collapse, several macroeconomic shocks occurred, exacerbating the uncertainty. Consumer sentiment in the US, as measured by the University of Michigan's index, fell to its lowest level in three years—a decline of about 6 percent to 50.3 points from October to November. This was significantly worse than forecasts, which had predicted only a slight drop to 53.2 points. Consumers' assessments of their current personal finances fell by 17 percent, while expectations of business conditions for the coming year dropped by 11 percent. This deterioration was not confined to a single demographic group but manifested itself "across the entire population, regardless of age, income, or political affiliation." The sole exception was consumers with larger stock holdings, whose sentiment rose by 11 percent—a clear indication of wealth inequality and the dependence of consumer sentiment on stock prices.

The shutdown effect: When the state paralyzes itself

The reason for this deterioration was closely linked to a structural failure: the US government shutdown, which entered its 38th day in November 2025, making it the longest shutdown in the country's history. A power struggle between Trump and the Democrats over the federal budget led to the paralysis of large parts of the US government. An estimated 670,000 government employees were furloughed, and another 730,000 worked without pay. The shutdown had direct economic consequences: it is estimated that a shutdown lasting four to eight weeks results in a permanent loss of $7 to $14 billion to the US economy—not just lost activity during the crisis, but permanent economic shortfalls that will never be filled.

Making decisions blindly: The paralysis of economic data

The shutdown also meant that the Bureau of Labor Statistics, the agency that publishes labor market and consumer price statistics, was deemed “non-essential” and ceased operations. This meant that key economic data—data on which the Federal Reserve bases its interest rate decisions—could not be released on time. This dramatically increased market uncertainty. Consumers feared not only their current economic situation but also the consequences of inflation: Expectations for the 12-month inflation rate rose to 4.7 percent, an increase from 4.6 percent the previous month.

Benefit from Xpert.Digital's extensive, fivefold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital

Xpert.Digital has in-depth knowledge of various industries. This allows us to develop tailor-made strategies that are tailored precisely to the requirements and challenges of your specific market segment. By continually analyzing market trends and following industry developments, we can act with foresight and offer innovative solutions. Through the combination of experience and knowledge, we generate added value and give our customers a decisive competitive advantage.

More about it here:

• Use the 5x expertise of Xpert.Digital in one package - starting at just €500/month

Voices of reason: The warning against the “enormous destruction of capital”

But the stock market turmoil of November 2025 wasn't solely the result of these macroeconomic factors. It was also the direct consequence of a fundamental reassessment of the risks in the AI ​​infrastructure boom. British hedge fund manager David Einhorn, who heads Soros Fund Management, put it succinctly: The figures currently circulating are "so extreme they are almost incomprehensible." He warned that the probability of "a massive destruction of capital in this investment cycle" is not negligible.

Suitable for:

• The dirty truth behind the AI battle of the economic giants: Germany's stable model versus America's risky tech bet

The Chinese disruption: When AI suddenly becomes cheap

This warning is reinforced by another development that has shaken confidence in the necessity of these massive infrastructure investments. The Chinese company DeepSeek demonstrated with its R1 and V3 models that impressive AI performance can be achieved at a fraction of the usual cost—not a marginal reduction, but a reduction of orders of magnitude. DeepSeek R1 costs about 2 percent of what users would pay for OpenAI's O1 model. Input costs $0.55 per million tokens for DeepSeek versus $15 for OpenAI. Output costs $2.19 for DeepSeek versus $60 for OpenAI. Even more remarkable, DeepSeek achieved this performance with just 200 employees and $10 million in development costs, while OpenAI employs 4,500 people and has raised $6 billion to date.

Suitable for:

• DeepSeek OCR is China's quiet triumph: How an open-source AI is undermining US dominance in chips

The efficiency shock: China's attack on the Western AI model

This is not a footnote about technological efficiencies. This is an existential threat to the entire justification system of the AI ​​infrastructure boom. If AI models can be developed with an order of magnitude less investment, then the $500 billion to $7 trillion investment plans are not forward-looking, but wasteful. The artificial scarcity that justified much of the valuation premium for Nvidia and other chipmakers—the assumption that only a few companies with the necessary capital could develop powerful AI—appears to be evaporating.

Real values, unreal assumptions: The energy problem of data centers

However, the situation is even more subtle in its macroeconomic complexity. Some experts argue that the AI ​​boom is fundamentally different from the dot-com bubble. While dot-com investments were primarily in "air"—business models that didn't work, companies that didn't generate real revenue—AI infrastructure investments lead to real, tangible assets: data centers, power supplies, physical hardware. A crash in this sector would undoubtedly cause massive pain—stock markets and commercial real estate would be severely impacted, gigantic data center projects would be sold off at rock-bottom prices, and hundreds of startups and service providers would go under. But at least for now, the damage would be limited in its macroeconomic significance, since the physical infrastructure itself retains some residual value.

The Achilles' heel of AI: The unquenchable thirst for electricity

However, this argument is undermined by energy challenges. Bain predicts that the additional global demand for computing power could climb to 200 gigawatts by 2030, half of it in the United States. The electricity consumption of AI data centers will increase from about 50 billion kilowatt-hours in 2023 to roughly 550 billion kilowatt-hours in 2030—an elevenfold increase. This means that the profitability of infrastructure depends not only on the ability to monetize AI services but also on the availability of sufficient energy sources at economical prices. If electricity costs rise or availability becomes a bottleneck—and both scenarios are likely—the entire equation collapses.

A structural vulnerability of the market

The US government shutdown also revealed another structural vulnerability: the markets' dependence on reliable economic data streams. When labor market statistics could not be released, uncertainty increased exponentially. This is a disconcerting scenario in modern economics. The Federal Reserve's monetary policy decisions rely on timely data on unemployment and inflation. When this data is unavailable, monetary policy becomes speculative. And speculative monetary policy in an already strained market environment leads to massive distortions. The fact that the shutdown lasted 38 days demonstrates deep institutional paralysis in the US that extends beyond mere partisan politics.

A door closes: The geopolitical dead end of AI expansion

Parallel to these developments, geopolitical shifts also took place that fundamentally challenged the AI ​​strategy. At a meeting between Donald Trump and Xi Jinping in South Korea, new trade agreements were negotiated. The US announced it would reduce its additional tariffs on Chinese goods from an average of 57 percent to 47 percent and maintain this tariff reduction until November 10, 2026. China committed to buying more soybeans from the US and suspending certain export controls on rare earth elements for one year. While this represented a de-escalation, it also acknowledged that the trade conflict was becoming detrimental to both sides.

The China Factor: A Cancelled Market and Its Consequences

The irony is that while the Trump administration reached a tariff extension with China, Nvidia made it clear that it had “no plans for the time being” to sell its new Blackwell chips in China and that it was “not actively engaging in discussions with China” because “it is up to China when they want to resume purchasing our products.” This statement by CEO Jensen Huang was more significant than any trade agreement. It meant that even if geopolitical tensions ease, US tech companies are unwilling to relax their export restrictions. This eliminates a massive potential source of demand from the AI ​​infrastructure calculations. The Chinese market, with its 1.4 billion people and its hunger for technology, remains closed to the Western AI infrastructure boom.

The rational response to an irrational assumption

The stock market reaction was therefore entirely rational, considering that hundreds of billions of dollars in investments in US AI infrastructure rest on the implicit assumption that this infrastructure would become the world's global AI data hub. If this market is inaccessible—and the signs indicate it won't be—then all the profit potential out of China must be eliminated. This isn't a minor margin hit, but the elimination of a quarter or more of the addressable market.

The end of the narrative: When good growth is no longer enough

In addition, valuation concerns intensified in the AI ​​sector. Palantir Technologies, the software company that develops analytical tools based on data analysis and machine learning, increased its revenue by 63 percent year over year in the third quarter and tripled its adjusted operating profit. But even with these growth results, which exceeded expectations, the stock fell by more than 5 percent after the release. Analysts pointed to the extremely high valuation. With a P/E ratio of 313, the stock is valued as if it would have to repeat its current earnings for 313 years to justify its market capitalization. This is not merely “ambitious” but mathematically absurd. The market began to realize that even rapid growth is not enough to support such valuations.

The rude awakening at Nasdaq

The Nasdaq index, which had risen 99.45 percent since 2020, signaled that the parabolic trajectory had to end sometime. The S&P 500, which had climbed nearly 95 percent since 2020, followed the Nasdaq down. For the first time in seven months, the tech sector experienced its steepest weekly loss. This wasn't a lapse by a few overvalued stocks, but a systemic failure of a narrative.

The asymmetry of risk: Who ultimately bears the losses?

The core question that arises is both simple and terrifying: If the AI ​​infrastructure investments don't turn out to be profitable—and all available evidence suggests they won't, at least not to the extent planned—who will be left with hundreds of billions of dollars in lost capital expenditures? The answer is complex because the investments have been concentrated in various positions. Nvidia, through its chip prices, has positioned itself at the bottom of the cash flow and will profit regardless of how profitable the data centers ultimately become. Meta and Microsoft, holding trillions of dollars in cash reserves, can presumably absorb the losses from operating profits. OpenAI, which isn't yet profitable, will either secure funding or fail. It is this asymmetry that makes the entire system fragile. The profits are concentrated among the hardware vendors, while the losses are diffused among the infrastructure investors.

The European perspective: A worried look across the Atlantic

The German economy, traditionally strong in mechanical engineering and industrial precision, is observing these dynamics with considerable concern. The dependence on US technology ecosystems for AI infrastructure is a structural problem that cannot be solved quickly. European companies are forced to either invest in US AI infrastructure or fall behind in technological capabilities. But the turbulence in the US market demonstrates that these investment decisions are being made on uncertain ground.

The end of uncritical euphoria?

Overall, the market reaction of November 2025 suggests not a mere correction, but a moment when markets began to recognize the structural weaknesses of the AI ​​infrastructure boom. Valuations were not defensive, profitability assumptions were not conservative, and geopolitical risks were not adequately priced. When these factors converged, the entire narrative collapsed. It remains to be seen how deep and how long this crash will last, but one thing is absolutely certain: the era of uncritical euphoria for AI infrastructure investments is over.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your national language!

Konrad Wolfenstein

I would be happy to serve you and my team as a personal advisor.

You can contact me by filling out the contact form or simply call me on +49 89 89 674 804 (Munich) . My email address is: wolfenstein ∂ xpert.digital

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitalization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Our global industry and business expertise in business development, sales and marketing - Image: Xpert.Digital

Industry focus: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More about it here:

• Xpert Business Hub

A topic hub with insights and expertise:

• Knowledge platform on the global and regional economy, innovation and industry-specific trends
• Collection of analyses, impulses and background information from our focus areas
• A place for expertise and information on current developments in business and technology
• Topic hub for companies that want to learn about markets, digitalization and industry innovations