Global technology companies are racing to pour trillions of dollars into artificial intelligence, from electronic chips and data centers to the development of smart models, but behind this massive spending, a fundamental dilemma emerges: can these investments generate profits commensurate with their costs, or are markets living in a bubble that could burst at any moment?
So far, the largest gains are concentrated among chip manufacturers, who are benefiting from the increasing demand for processors, while other companies continue to spend billions of dollars to develop and operate the technology without proving their ability to generate sustainable returns.
This disparity raises growing fears of a repeat of the dot-com bubble scenario of the late nineties, when investors rushed behind future promises before technology stocks collapsed and trillions of dollars in market value evaporated.
From a tech boom to a potential bubble
Current concerns do not mean that artificial intelligence is a valueless technology; rather, the problem lies in the possibility of overestimating its future profits compared to actual returns.
The dot-com bubble saw the rise of companies that received massive funding despite weak revenues and the absence of business models capable of generating profits. When investor confidence waned, valuations collapsed, and the Nasdaq index lost about 77% of its value between its peak in March 2000 and October 2002.
Today, the question is renewed as to whether the massive investments in artificial intelligence are based on economic demand capable of generating real profits, or on optimistic expectations that may not materialize as quickly as the markets are betting.
Massive spending and uncertain profits
The current boom differs from the dot-com bubble in that a large part of the spending is driven by giant companies that have existing businesses and significant financial resources, such as Microsoft, Alphabet, and Meta, alongside Nvidia, which is benefiting from the demand for chips.
However, the strength of these companies does not eliminate the risks, especially with the rising costs of building data centers, purchasing processors, and providing electricity and cooling, in addition to the need to constantly update equipment with the rapid development of technology.
Concerns increase when revenue growth coincides with continued losses and borrowing. CoreWeave provides an example of this, as it expanded into data centers by relying on financing, amid expectations of revenues of nearly $13 billion during 2026, against losses estimated at about $2.13 billion, according to the figures mentioned in the reports.
This case illustrates that high demand does not guarantee profitability, especially when the costs of expansion and financing exceed the companies' ability to generate liquidity.
Intertwined financing doubles the risks
The risks are not limited to a specific company, but extend to the financial relationships between technology companies, service providers, and investors.
Some companies finance others through investments or loans in exchange for future commitments to purchase their services and products. If one party falters, expected revenues may decline, and the value of investments and debts held by related parties may be damaged.
According to the figures in the article, the International Monetary Fund, in a review covering 14 deals since 2023, tracked equity investments of about $300 billion and debts of $650 billion, against commitments for future revenues of $1.4 trillion. These commitments do not represent realized revenues, but rather expected cash flows that remain tied to the parties' ability to fulfill their obligations.
Who bears the cost of the boom?
Achieving sustainable economic returns represents the most important test for the future of artificial intelligence. The technology's success in attracting users is not enough on its own, unless companies can increase their revenues or reduce their costs to cover the massive investments.
Companies face an additional challenge represented by high capital expenditure, the possibility of processors becoming obsolete before their cost is recovered, as well as constraints related to providing electricity and expanding the energy networks necessary to operate data centers.
In light of these challenges, some companies may find themselves forced to scale back expansion plans or borrow larger amounts if returns fall short of investor expectations.
Is the burst approaching?
The International Monetary Fund warns that weak expected returns on artificial intelligence investments could lead to a correction in asset prices, with the potential for losses to be amplified by borrowing and intertwined financial relationships.
The impact of any sharp decline could extend to investors, investment funds, and global markets, especially if the decline in technology company valuations coincides with rising financing costs and declining confidence.
However, the possibility of an investment bubble bursting does not mean the disappearance of artificial intelligence or the failure of the technology itself. The internet experience proved that innovation can continue to grow even after the collapse of companies that the markets had overvalued.
In the end, the future of the artificial intelligence boom will not depend on the amount of money poured into it, but on the ability of companies to turn these investments into sustainable profits and cash flows. Between technical promises and financial reality, the crucial question remains: will profits outpace investor fears, or will the expectations gap be exposed before the trillions of dollars invested are recovered?
