The AI bubble is becoming a major threat to the American and global economies
Prof. V Yu Katasonv, Sovetskaya Rossiya, December 26, 2025 —
In English-language media, especially in the American media, the “AI bubble” became a key topic this fall. This refers to the stock market bubble that arose as a result of the rapid development of so-called “artificial intelligence” (AI) in the United States and several other countries.
Artificial intelligence, in its broadest sense, is defined as a set of machine (primarily computer) tools capable of performing tasks comparable to human intelligence (such as perception, learning, reasoning, problem solving, and decision making). British mathematician, logician, and cryptographer Alan Turing (1912–1954) was the first person to conduct extensive research in the field, which he called machine intelligence. Artificial intelligence was formally recognized as an academic discipline in 1956; its founders were the American mathematicians, logicians, and engineers John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon.
Public interest in AI surged at the turn of the last decade, with companies, universities, and laboratories, overwhelmingly based in the United States, pioneering significant advances in artificial intelligence. Some of the most prominent uses of AI include advanced search engines such as Google Search, Bing, and Yandex; recommendation systems used by YouTube, Amazon, and Netflix; human-language interactions such as Google Assistant, Siri, Alexa, and Alice; autonomous vehicles such as Waymo; generative and creative tools such as ChatGPT and Apple Intelligence; and analysis in strategy games such as chess and Go.
Forbes magazine ranked the top 50 companies in the world specializing in AI for 2024. The ranking was based on the amount of investment in AI development. Only 11 companies on the list were not American (two British, two French, two Canadian, two Dutch; one each from Australia, Germany, and Sweden). The remaining 41 companies were from the United States.
The more specific areas of focus for the AI companies on the Forbes list included: AI model development; neural network development; AI application software; data storage and analytics; enterprise search engines; image generation services; drug discovery and development; video generation services; avatar and video generators; industrial machine maintenance, etc.
Until recently, there was a certain euphoria surrounding AI in the media. Business representatives (especially those involved in AI and IT), many politicians, and journalists claimed that AI would usher in a true revolution in all spheres of society. AI would extend human lifespan and make it more efficient, increase economic productivity, ensure national and individual security, and so on. In short, AI was presented as a gateway to a brighter future.
Of course, there were (and still are) many opponents of such unbridled optimism. Every new direction in science and technology not only creates opportunities to solve various socioeconomic problems but also poses serious new threats. This is roughly what happened with nuclear energy. Its first practical application led to the deaths of hundreds of thousands of people (the atomic bombs dropped by the Americans on Hiroshima and Nagasaki). You can even find a dedicated website online that maintains a registry of AI risks, the “AI Risk Repository.” This registry currently lists 777 risks.
But now we’re not talking about the direct risks of AI itself, but rather the risks of the “bubble” that AI companies have created in the stock market. After a fairly long period of AI-fueled euphoria, early fall of this year saw talk of a stock market bubble. The media is increasingly reporting figures indicating that investment and expectations around AI are growing faster than the technology’s actual capabilities. IT companies’ investments in AI are growing, and their market capitalization is growing even faster. Consequently, investor capital inflows into IT companies are growing. Meanwhile, IT companies are either generating little or no profit.
Just a few companies are making a significant contribution to the growth of AI investments. In 2023, BofA analyst Michael Hartnett even coined the name “Magnificent Seven” for the leading AI companies. It includes the following American tech companies: Apple, Amazon, Alphabet, Nvidia, Tesla, Meta (designated extremist and banned in Russia), and Microsoft. These companies are engaged in AI R&D and implementation, while Nvidia is a leader in the production of chips for related equipment. This spring, the Mag7 group achieved truly dizzying success in terms of stock prices.
In April of this year, Nvidia became the first global company to surpass $5 trillion in market capitalization; even after a December rebound to $4.4 trillion, it remains the world’s largest company and remains the world’s number one.
Here are the market capitalizations of the other Magnificent Seven companies (in trillion dollars): Apple – 4.0; Alphabet Inc. – 3.6; Microsoft – 3.5; Amazon – 2.3; Meta – 1.5; Tesla – 1.3 trillion. Over the past five years, the Mag7 companies’ market capitalization has grown 3-7 times. Meanwhile, the S&P 500 index has grown by only 44%, with half of this increase due to the Mag7.
Experts estimate that the market price of American tech companies in the third quarter of 2025 exceeded their earnings by 41 times. This is the price-to-earnings ratio (P/E); economics textbooks typically state that a P/E ratio of less than one is a sign of trouble. A company with a P/E ratio of 1 to 2 should be considered healthy and stable. Anything higher may indicate a risk of a bubble forming. More precisely, the risk of this bubble bursting. Around September, concerns about an AI bubble and its potential collapse began to grow in America.
A story from more than a quarter-century ago inevitably came to mind: the dot-com crash on the NASDAQ stock exchange, where high-tech companies trade. In the late 1990s, high-tech companies boomed on NASDAQ and other stock markets. But back then, it was a boom driven by expectations for a technology like the internet. Just as we see euphoria and expectations for a “bright future” from AI today, so too was there euphoria about the internet and new communications technologies (expectations of increased productivity, the creation of new markets, increased employment, etc.).
From June 1999 to March 2000 (nine months), the NASDAQ-100 index rose to 4,800 points, or by 130%. By the beginning of 2000, the P/S ratio (the ratio of a company’s market capitalization to its revenue) for the top 100 companies on the NASDAQ market approached 25. This was already the situation that became known as the “dot-com bubble” (dot-com is a term that came to be applied to a company whose business model was based on operating within the Internet). The dot-com bubble burst on March 10, 2000. It is said that before the bubble burst, the P/S ratio for all 100 NASDAQ companies exceeded 33. And the P/E ratio at that time was 43. Let me remind you that at the end of September of this year, the P/E ratio reached 41.
After the bubble burst, the NASDAQ index fell by nearly 80%, with the telecom (dot-com) sector losing nearly 90%. As a result, confidence in venture capital investments was undermined. Hundreds of internet companies went bankrupt, were liquidated, or sold. Many related industries, such as advertising and logistics, curtailed their operations due to falling demand for services.
The entire US economy was rocked, falling into recession in the first quarter of 2000. However, the country’s GDP growth remained positive for the entire year.
Let’s return to America in 2025. Many are alarmed not only by the steep market capitalization dynamics of the “Magnificent Seven,” but also by the fact that AI business leaders are truly generous with their massive investments in AI projects. These investments are being used to build data centers, purchase GPUs (graphics processing units), create various infrastructure, and so on. According to Gartner, Inc., global spending on artificial intelligence (AI) in 2022 was under $70 billion; in 2025, it is estimated to reach nearly $1.5 trillion. Next year, spending is projected to reach $2 trillion, or 32 times more than in 2022. The bulk of these investments will be directed toward the Mag7.
Such gigantic investments have previously been made in nuclear energy or space exploration. But there’s no guarantee that the gigantic physical assets currently being created in AI will ever generate returns comparable to those of nuclear and space projects. In the event of a collapse, all these gigantic assets will become worthless junk.
Not only independent experts, but also US government officials and even executives of AI companies are talking about the AI sector approaching a red line. Ray Dalio, director of Bridgewater Associates, noted earlier this year that the current level of investment in AI is “very reminiscent” of the dot-com bubble.
Julian Garran, an analyst at the research firm Macrostrategy Partnership, made an even harsher statement in October. He said the AI sector has become the largest speculative bubble in human history, 17 times larger than the dot-com bubble. The dot-com bubble didn’t extend beyond the US economy. The current AI bubble, however, looms large over the entire global economy and could trigger a global crisis.
Sam Altman, CEO of the renowned US AI company OpenAI and creator of ChatGPT (a chatbot with generative artificial intelligence), has stated that he believes the AI bubble is already inflating.
The International Monetary Fund (IMF) has also addressed the issue of the “AI bubble.” IMF Chief Economist Pierre-Olivier Gourincaz expressed concern on behalf of the Fund that the “tech bubble” created by significant investments in AI could burst. An IMF report published in October stated: “The potential collapse of the AI boom could rival the dot-com crash of 2000-2001 in severity, particularly given the dominance of a few technology companies in market indices and the involvement of less-regulated private debt financing much of the industry’s expansion.”
In November 2025, Bank of America (BofA) analysts surveyed more than two hundred global portfolio managers managing $550 billion. Every second respondent named the inflating AI bubble in the stock market as the top threat to the US and global economies over the next year. In the previous survey, conducted in August, the top threats were identified as the risks of a major trade war (provoked by US President Donald Trump), accelerating inflation, and a Federal Reserve rate hike. Back then, only 11% of respondents feared an inflating AI bubble in the stock market. And fears of an AI bubble ranked only fifth on the list of anticipated threats and risks. Now, they have risen to the top spot.
However, there are optimists among experts who believe that the “AI bubble” in the US stock market will not burst within the next year. Various reasons are cited for this, but one is truly compelling. Jerome Powell, head of the US Federal Reserve, announced that the quantitative easing (QE) implemented by the US central bank will end on December 1, 2025. He hinted that a new phase of quantitative easing (QE) will begin next year. This means that the Fed’s “printing press” will likely be revved up again, and the key interest rate will be lowered. Well, thanks to the QE policy, a stock market crash may not occur. But this does not eliminate the threat, but merely delays it. Large amounts of new, cheap money will further inflate the “AI bubble.” Therefore, a delayed crash will be even more severe and destructive.