Artificial intelligence is the number one topic in business today. Larger and smaller companies around the world are investing substantial sums in AI projects and have high expectations regarding returns. However, many of these expectations remain unmet.
This frustration is also at the heart of recent insights that Jensen Huang presented, CEO of NVIDIA, at an AI conference. According to him, demanding an immediate return from artificial intelligence is like forcing a child to create a business plan. He therefore adds that the most important thing is to persevere and test different approaches.
The numbers speak for themselves
Research figures also illustrate that companies are dissatisfied with the results. According to one study, 88% of organizations use AI in at least one segment of their business. According to an MIT study from August 2025, 95% of generative AI initiatives end in failure or fail to generate a clear profit.
Only a small percentage of AI projects achieve the expected return, and even fewer exceed the cost of capital. This means that most of them end without the expected profit, increasing frustration among executives and investors alike.
Perseverance is essential
“Let a thousand flowers bloom, and only then choose the most fruitful ones”, Huang said at the conference, pointing out that the most important thing is to persevere and observe what actually works. Only by experimenting with models and tools can their true long-term value become apparent. He is thus urging companies not to abandon AI projects at the first signs of slow returns, but to give them enough time.
According to experts, expectations of a clear profit are assessed over too short a period. Huang's position is therefore unequivocal: AI should not be evaluated according to short-term financial indicators, but according to its gradual integration into corporate processes.
Only a long-term perspective will reveal the true potential of technologies that are just beginning to transform the shape of business today. And this approach is proving to be extremely important.
According to a Deloitte report, it takes most companies 2 to 4 years before they begin to see a satisfactory return on their AI investment, which is significantly longer than is typical for other technology investments. Only 6% of companies report seeing a direct return on investment in less than one year.
Strategy may also be to blame
The fact that many companies entered the AI era without a clearly defined strategy also contributes to returns that fall short of their expectations. In practice, this means that they often integrated the technology into processes in isolation, for example only in marketing, customer support, or HR. The result is not dramatic change, but incremental improvements that are more difficult to reflect in the overall figures.
Another factor is the indirect costs of operating AI. In addition to investing in tools, companies must also account for infrastructure expenses, security, and employee training. And people play a crucial role here. Without the right approach to AI, its potential remains untapped.
The right approach delivers results
None of this means that investments in artificial intelligence should not deliver results. The proof lies in companies that persevered and integrated AI into their processes consistently and gradually. Today, these companies report not only increased productivity, but also faster innovation. But although the original expectations may have been set in terms of months, practice shows that the actual timeframe is measured in years.
According to the data and analyses, AI is simply not a quick financial shortcut, but a long-term strategic investment. And those who can overcome the initial period without immediate results may gain a significant competitive advantage in the future.
