According to an analysis by Goldman Sachs, the large financial investments made in artificial intelligence by technology companies may require several years to start generating profits. Before this period, these companies must reach the break-even point, the moment when the revenue obtained is sufficient to cover the costs of the investments made.
Strategist Ryan Hammond estimates that cloud computing giants, such as Amazon, Oracle, and Microsoft, will need to generate approximately US$ 300 billion (equivalent to BRL 1.55 trillion) from AI in the coming years to reach this break-even phase.
It is observed that the cloud computing revenues of hyperscalers showed acceleration this year. In the second quarter of 2026, the annualized rate exceeded the previous trend for AI expansion by about US$ 70 billion (BRL 364 billion).
Already announced future contracts from the group exceed the mark of US$ 1.5 trillion (BRL 7.8 trillion). However, for Hammond, these figures do not yet indicate that the investments are close to paying off.
The strategist emphasized that it is expected that AI users will spend about US$ 1 trillion (BRL 5.19 trillion) annually on AI applications for hyperscalers to be able to generate solid returns on invested capital. This projection is conditional on a significant increase in spending on AI applications, which, according to Hammond, would also help the application layer achieve robust profit margins relative to computing costs.
This warning comes at a time when investors are resuming interest in the shares of large technology corporations, motivated by optimism regarding AI. The analysis points out that a crucial step before large investments begin to yield consistently is converting the growing demand for computing and applications into revenue capable of sustaining the inherent costs of the process.