AI is rapidly entering the stock market. What could possibly go wrong

The main thing in the material:

— SpaceX has launched the largest wave of technology IPOs since the dot-com era, and OpenAI and Anthropic may go public this fall.

— The main risk for investors is not related to the revolution of technology, but to the ability of AI companies to turn growth into sustainable profits.

— Increased competition and a possible price war can put pressure on developers' margins even before they reach a stable cash flow.

— IPOs will test not only the demand for AI assets, but also the market's faith in the long-term economy of artificial intelligence.

The world is preparing for the largest wave of technology IPOs since the dot-com era. SpaceX has already gone public, OpenAI and Anthropic are also preparing for a public offering. Investors are rushing to take seats in the front row of the new technological revolution, fearing to repeat the mistake of those who once missed Amazon, Google or Nvidia. However, history has already shown that companies that change the world are not always the best investment solution. Izvestia investigated what risks the market may underestimate now.

Do not repeat the fate of Cisco

Experts are increasingly associating the excitement around AI with the dotcom era in the late 1990s. Although, according to analysts, stock valuations and investor enthusiasm have not yet reached the peaks seen at the turn of the century.

The S&P 500 index has grown by more than 50% this year from the lows of 2022, the Nasdaq Composite - by more than 70%. The share of the technology sector in the capitalization of the S&P 500 has reached 32%, the highest since the dotcom bubble. At the same time, only three companies — Microsoft, Apple and Nvidia — occupy more than 20% of the index. So, with a capitalization of about $5 trillion, Nvidia today ranks first among public companies in the world. Over the past five years, the chipmaker's shares have grown by almost 4,300%. Cisco securities, one of the symbols of the Internet revolution along with Dell, Microsoft and Intel, have risen in price by about the same amount over the same period.

In 2000, Cisco briefly became the most expensive company in the world with a capitalization of more than $550 billion, surpassing even Microsoft. Investors were confident that Cisco would be at the center of the formation of a new economy. She supplied network equipment that was needed by Internet companies. When the dot-com bubble burst, the Nasdaq Composite index, after prolonged growth, collapsed by almost 80% from its peak in March 2000, the S&P 500 index, which had doubled before, lost almost half of its value. Cisco shares then collapsed by more than 80%. But at the same time, unlike many Internet companies of that era, Cisco did not disappear and continued to develop. Nevertheless, investors had to wait more than 25 years before the company's securities could return to peak levels. In December 2025, Cisco shares were above the historical maximum for the first time, at $80.25.

The Cisco story has shown that the market can absolutely correctly assess the importance of technology — the Internet has really changed the world — and at the same time make a mistake in assessing the investment potential of leading companies.

Today, new players are entering the public market, offering solutions to the world that can define the next economic era.

Investors buy the promise

The first large-scale test of investors' willingness to pay for the future economy was the SpaceX IPO. Shortly before going public, Elon Musk's company changed the investment narrative, supplementing the space business with a story about the future of artificial intelligence. SpaceX, after merging with the AI company xAI and the deal to purchase the Cursor startup, began to position itself as a new infrastructure player in the artificial intelligence market. In the issue prospectus, the company estimated the potential market for its AI areas at $26.5 trillion (this is more than 90% of the total potential market of the company), announcing plans to build data centers in space, as well as earn money by renting its computing power (such a deal already exists with Anthropic).

SpaceX sought to enter the market ahead of OpenAI and Anthropic, hoping to be the first to take advantage of the growing investor demand for AI assets. As a result, the company raised $75 billion, its market capitalization after the placement exceeded $2 trillion, and Musk became the world's first trillionaire. The IPO showed that investors are willing to invest in companies that promise to make money on AI long before the economics of this business become fully understood by the market.

One of the main unpleasant consequences of the IPO for SpaceX will be the need to live by the rules of the public market, that is, to disclose financial statements quarterly and answer questions that may arise on Wall Street. Before the listing, Musk could afford to invest tens of billions of dollars in projects with a long payback horizon. After going public, the company will have to implement a long-term strategy and at the same time meet the expectations of investors. Any delays in the development of new products, rising costs, or weak financial performance will immediately affect the stock price.

Formally, SpaceX remains unprofitable. In 2025, the company generated almost $19 billion in revenue (+33% compared to 2024) and posted a net loss of $5 billion; in the first quarter of 2026, revenue was $4.7 billion and net loss was almost $4.3 billion. The main source of profit remains the telecommunications business, Starlink, which generated about $1.2 billion in the quarter. At the same time, the space direction and the AI division suffered financial losses. Last year, xAI generated $3.2 billion in revenue, but the operating loss reached almost $6.4 billion. Capital expenditures totaled $12.7 billion, three times higher than the rocket business. The funds were allocated primarily for the construction of Colossus data centers in Tennessee.

Despite this, SpaceX's fundamental business model looks sustainable. The company has an operating business with clear revenue sources, scalable infrastructure, and long-term contracts. The losses in this case are mainly related to aggressive investments in new areas. Therefore, the main question today is not related to the viability of SpaceX, but to how expensive the market is willing to evaluate its future.

The successful listing of the company has become an important landmark for OpenAI and Anthropic. However, unlike SpaceX, the investment case of these companies is mainly based on expectations of the future growth of the artificial intelligence market.

Make a profit

Earlier this month, Anthropic, which launched the Claude AI model, and OpenAI, which develops ChatGPT, filed IPO documents with the Securities and Exchange Commission. The listing of both companies may take place in the fall. When entering the public market, OpenAI plans to reach a capitalization of $1 trillion at a current valuation of $852 billion.

According to forecasts made by OpenAI to investors before the next round of financing, by 2028 the company expects to spend about $121 billion per year on computing power alone to develop new artificial intelligence models. Such a scale of expenditure would be unprecedented even by the standards of the largest publicly traded companies.

According to Anthropic, which is estimated at $965 billion, much less financial data is available. The company's revenue is expected to grow to $10.9 billion in the second quarter of this year, compared to $4.8 billion in the first. If the forecast is correct, Anthropic will show an operating profit of $559 million for the quarter for the first time. Its quarterly revenues are growing faster than Zoom's during the pandemic or Google and Facebook (owned by Meta, recognized as extremist and banned in Russia) on the eve of the IPO.

Despite this, the market still has virtually no access to data on the cost structure and scale of losses - until the publication of a full—fledged issue prospectus, Anthropic remains one of the most expensive and at the same time the least transparent tech companies among those going public.

Both companies are showing incredible growth rates, but their business requires unprecedented investments in computing infrastructure. Investors will have to figure out how much of this growth will translate into free cash flow after paying for model training, chips, data centers, and energy costs. So OpenAI and Anthropic will have to prove not only their technological leadership, but also their ability to turn artificial intelligence into a high-margin industry.

Price war instead of monopoly

Even if companies continue to grow revenue rapidly, this does not guarantee that they will remain highly profitable in the future. There are enough players in the AI market — OpenAI, Anthropic, Google, Meta, xAI, DeepSeek and other developers are simultaneously releasing more and more powerful models. The technological gap between them is narrowing. More and more companies are starting to use several models at once, automatically distributing tasks between cheap and expensive tools depending on the complexity and cost of the request. This approach makes it possible in some cases to reduce AI costs by 95%.

It turns out that corporate clients are receiving more and more cost optimization solutions and are becoming less dependent on a specific model supplier. For OpenAI and Anthropic, this means increasing pressure on prices. OpenAI is already considering the possibility of significantly reducing the cost of its products in anticipation of similar steps by Anthropic. And this is an alarming signal for investors. A price war can lead to lower developer margins and make it more difficult to achieve sustainable profits. This means that investors are at risk of buying not the future Google, but players in the market with huge demand, high costs and constant pressure on prices.

That is why the IPOs of OpenAI and Anthropic will test not only the demand for artificial intelligence, but also investors' faith in the long-term economics of AI. If expectations are met, artificial intelligence will become the main investment topic in the coming years. If not, the market will have to face a painful reassessment of its bets on one of the most ambitious stories of the modern capital market.