Is the boom in artificial intelligence investment coming to an end?

Optimism around investing in AI stocks is waning, and fears and doubts about the technology revolution are growing louder among analysts. Does this mean the end of the AI boom? "For investors, the biggest problem is not the prospects of quick or distant returns. The problem lies in the threefold uncertainty with which investors' expectations collide: firstly, the uncertainty of effects, secondly, the uncertainty of regulations, and thirdly, the uncertainty related to the right business decisions," emphasises Radosław Jodko, an investment expert.

No investment has been talked about so often and with such optimism this year. Artificial intelligence companies were of great interest to investors until... they stopped somewhat. Analysts are getting louder and louder about the fact that artificial intelligence is not delivering the expected returns, considering both the effects and the investment.

Bank of America even writes in its report that it is time for artificial intelligence to move from the level of promises to a real demonstration of how it works and what it achieves, while ‘any discrepancy between investment and revenue generation will come under increased scrutiny’.

Data centre development

Most anticipated by analysts today are Nvidia's Q2 results, which the company is expected to release on 28 August. Nvidia is a maker of GPUs, used by AI technologies, whose shares have tripled over the course of this year, only to score steep declines recently. Analysts' expectations are high, predicting revenue growth of 112 per cent, but the question of whether Nvidia is in the midst of a speculative bubble is increasingly being heard.

"Nvidia is a good example of the uncertainty we face with technology companies. The volatility of the technology sector generates huge spikes that are influenced by so many factors that are difficult to calculate, such as whether another brilliant chipmaker will accidentally emerge, what the regulations will look like and so on. Today, data centres are developing, so we used to think that it is - generally speaking - semiconductor companies that perform better than software companies. And this is generally true, although we are seeing surprises here too, as this year this trend was broken by a data analysis software producer and PLTR's shares jumped by 86 per cent," Radosław Jodko, investment expert at RRJ Group, points out.

The importance of data centres can be clearly seen in the sales results of, for example, the aforementioned Nvidia. The previous quarterly report showed that the data centre division was responsible for a 92 per cent year-on-year increase in sales.

Does business not believe in AI?

"Certainly, investors have reached the point where, beyond promises and expectations, they want to see results. And not just the effects of the development of generative AI, because we are obviously seeing those. But are we dealing with an effectively executed monetisation of AI? With a cascading process where more companies in more industries investing in AI change their own business models? We don't hear too much about such, and if we look at reports from companies conducting this kind of analysis, such as the McKinsey reports, there is not much optimism in companies and the wider business about the revolutionary changes that AI would make in business. Instead, we see that investments in AI are costly and uncertain because they do not necessarily translate into profits for companies resulting from such investments. The optimisation resulting from AI implementation alone is simply not enough," argues Jodko.

The problem of monetising AI

Is AI therefore overrated? "I would not be so fundamental. We are seeing a lot of successful projects and certainly a lot more interesting ones will find their followers among investors as well. But we will only see a real boom when companies outside the technology sector manage to implement AI so successfully that it has a knock-on effect on their own businesses and industries. In the meantime, we are still at the stage of finding answers to much earlier questions, such as how quickly technology companies will accelerate pilot programmes to commercial implementation and how to solve the AI monetisation problem. No one has still found a good answer to these," concludes Jodko.

Recent market analyses indicate that most enterprise software vendors will not be monetising generative artificial intelligence or conversational AI by the end of 2025.