Artificial intelligence has enormous potential to transform the global economy, but investors and businesses should be patient. As recent macroeconomic studies show, the noticeable impact of this technology on labor productivity will only become tangible in a few years, and the adaptation process itself will be accompanied by high market volatility.
Implementation requires time, not just capital
In a recent analysis, I concluded that we are witnessing a unique situation. In the entire history of observing technological cycles, I have never encountered an innovation that promises such a radical leap in efficiency as AI. However, the key nuance is that the corporate sector will need years to fully integrate these tools into operational processes. Only then can we talk about "maximum returns."
This means that the current hype around AI company stocks, which we have observed, will inevitably lead to a correction. The long-term effect, however, will be much broader than any market cycle. The world is entering a period of adaptation that will be accompanied by a reassessment of expectations and, likely, significant price fluctuations.
Labor market and inflation: new risks
Despite fears of mass automation, I share the view that the development of large language models (LLMs) will not lead to a collapse in employment. History teaches us that every technological breakthrough has sparked similar concerns, but ultimately only transformed the labor market, rather than shrinking it. AI will become a powerful tool for enhancing employee efficiency, not a straightforward replacement for them.
Special attention should be paid to AI's impact on inflation. An analysis of 250-300 years of technology adoption reveals a paradox: increased productivity alone did not lead to sustained price declines. On the contrary, the growth in business efficiency driven by AI may be accompanied by additional inflationary risks related to restructuring supply chains and rising demand for computing resources.
Investment bubble or fundamental shift?
Capital expenditures by the largest technology giants on AI infrastructure this year, in my estimation, will reach $527 billion. That is $62 billion more than was forecast at the start of the earnings season. However, the scale of investment still significantly outpaces the measurable economic impact.
According to recent calculations, in 2026 these investments will add only about 0.3 percentage points to real GDP growth and just 0.1 percentage points to measurable growth. This suggests that we are in the early, costly stage of the cycle, where infrastructure spending is growing faster than the returns from implementation.
Analyst's perspective
The market is already pricing in future super-profits from AI, but the real economy is not yet ready for such speed. I expect that over the next 12-18 months, we will see a period of "disappointment," as companies begin to report modest results from AI adoption. However, it is precisely at this moment that the best entry points will open up for long-term investors who understand that the real effect will only arrive after 2027-2028.