At the beginning of November, global markets seriously questioned for the first time whether the artificial intelligence wave had outgrown the reality it could deliver.
After almost two years of continuous growth, the most valuable technology companies experienced a sharp decline.
On 5 November, the Nasdaq fell by more than 2%, the S&P 500 dropped by more than 1%, and some stocks lost up to 8% in a single day.
This was the first sign that confidence in the "AI mania" was cooling down.
Nvidia, whose chips are at the heart of every major AI platform, became the symbol of this change.
At the end of October, Nvidia became the first company in history to reach a market value of around $5 trillion, just three months after surpassing the $4 trillion threshold.
When investors began to withdraw capital, its shares fell by about 7%, and the rest of the sector followed.
The decline was not caused by poor business results but by growing concerns that the market depends on a small number of companies and on profit expectations that have yet to be realised.
A global correction
On Asian stock markets, the consequences were even more pronounced. Japan's Nikkei 225 fell by 4.7%, South Korea's KOSPI by more than 6%, and leading semiconductor companies such as Samsung and SK Hynix lost between 8% and 10% of their value.
This made it clear that the correction was not local but global – the fall in confidence in the AI industry has affected the entire technology chain.
In the first week of November alone, almost 1,000 billion dollars of market value was wiped from companies whose business is directly related to AI
Financial data illustrate the scale of this trend. In the first week of November alone, almost 1,000 billion dollars of market value was wiped from companies whose business is directly related to artificial intelligence.
The Financial Times described the period as "the worst week for tech stocks since April."
Reuters notes that the technology sector now accounts for more than 35% of the value of the S&P 500 index – a concentration not seen even during the dot-com era.
The “third bubble” of the world economy
What is the essence of the problem?
Companies are investing vast sums in infrastructure, but concrete results have yet to materialise.
According to industry data, the four largest firms – Amazon, Google, Meta, and Microsoft will spend more than $300 billion this year alone on developing AI systems and data centres.
OpenAI, the symbol of the current generation of artificial intelligence, achieved revenue of 4.3 billion dollars in the first half of 2025 but also recorded a loss of 13.5 billion.
Such a gap between expectations and actual results is making investors nervous, especially those who remember how similar stories ended twenty years ago.
No one is talking about a system crash yet, but few claim that nothing will happen
Bankers and analysts who until recently spoke of a "revolution" now use the word "correction."
The president of the World Economic Forum, Børge Brende, has warned that AI, along with cryptocurrencies and rising government debt, could become the "third bubble" of the world economy.
Experienced investor Mark Mobius (founder of the Mobius Capital Partners fund, specialising in developing markets and a long-time analyst of global financial trends) estimates that some leading AI stocks could lose up to 40% of their value if indicators do not improve.
No one is talking about a system crash yet, but few claim that nothing will happen.
The end of euphoria
The question now is not whether investments will decrease, but how their structure will change.
Companies that have spent billions on chips, servers and data centres will need to demonstrate tangible benefits – either through increased productivity or new revenue streams.
Otherwise, investors will seek other sectors. Pension and insurance funds are already announcing a reduction in exposure to "AI giants" and a shift towards industries that offer more stable returns.
Is this the end of AI optimism or the beginning of a more realistic phase?
A broader question arises: is this the end of AI optimism or the beginning of a more realistic phase?
Experience with previous technology cycles shows that no innovation disappears because of a stock market crash.
The Internet survived the dot-com crisis – the collapse of the technology market in the early 2000s, when thousands of Internet companies failed after their share values plummeted due to inflated expectations about earnings in the digital economy – and social networks survived data scandals, so AI will continue to evolve.
The difference lies in the pace. The market is now correcting unrealistic expectations and seeking a balance between opportunity and value. This does not necessarily mean the end of growth, but it does mark the end of euphoria.
Where are the concrete results?
A political dimension is also unfolding in the background. Governments that have invested billions in AI strategies now face the same question as investors: where are the concrete results?
Tax treatment for AI companies and subsidies for domestic chip production are already being debated in Washington, while the European Union is considering how to prevent a market from overheating – a situation it has partly fuelled with subsidies and preferential loans.
Should the downward trend persist, the emphasis will shift from innovation to fiscal responsibility.
Governments will have to explain how their investments in AI are truly justified and financially sustainable.
For the first time in many years, the very definition of technological progress is being re-examined
The decline in AI stocks has a deeper consequence: for the first time in many years, the very definition of technological progress is being re-examined.
The market is accustomed to measuring each new phase of the digital industry by its speed of growth rather than its ability to deliver real value to society and the economy.
Investors are now realising that AI may be a revolutionary technology, but its contribution to productivity remains unmeasurable. In the United States, productivity growth in 2025 is less than 1.5%, although last year's projections suggested it would reach 2.5% to 2.7% due to the effects of artificial intelligence.
From hype to sustainable models
At the micro level, the issue is similar. Most large companies are introducing AI tools into their operations, but most employees make limited use of them.
A McKinsey Global Institute survey from October shows that only 11% of companies using generative artificial intelligence report a real increase in efficiency.
Instead of a revolution, a gradual adjustment is underway – and the market is now recognising this
This is far from the expectation that AI would "replace half the jobs." Instead of a revolution, a gradual adjustment is underway – and the market is now recognising this.
Another signal of change comes from investors in Asia and the Middle East, who until now have been the main source of capital for AI startups.
Funds from Saudi Arabia and the United Arab Emirates are reducing new investments in American projects, redirecting part of their capital, and focusing on their own AI initiatives.
This weakens the flow of foreign capital that has fuelled the sector's growth for years.
China, while still investing in AI, is shifting its focus to domestic applications in industry and manufacturing rather than generative models similar to ChatGPT.
All this means that artificial intelligence is entering a phase where it will no longer grow on the basis of promise but on the basis of usability.
In the upcoming year, it will become evident who has a truly sustainable model and who has simply ridden the hype.
It will be a moment of separation between those who develop the technology and those who sell its story.
For the global economy, this is a sobering period. If AI is truly the "new electricity," as many have called it, it must now demonstrate the same qualities: stability, predictability, and measurable results.
Until that happens, markets will react nervously and capital will seek safer havens. This is where the real value of the crisis becomes apparent – it does not signal the end, but a return to reality.
