Artificial intelligence has become one of the decade's biggest investment trends: major tech companies are putting unprecedented sums into data centers, specialised chips, and next-generation AI models. Analysts at XTB, a global investment application available in Hungary, examined whether these record investments can deliver the expected productivity gains and business value for firms.
Shift toward "good enough" models and cost efficiency
According to XTB's latest market analysis, the industry has entered a new phase: companies are less focused on deploying the most advanced—and most expensive—AI systems and are increasingly turning to cheaper, "good enough" models that provide adequate performance at substantially lower cost. This trend does not necessarily mean reduced AI usage overall; rather, there is a shift away from costly "frontier" models toward more cost-effective solutions. At the same time, many AI developers still lack a clear path to monetise their infrastructure and compute investments.
Scale of spending and investor expectations
XTB's experts estimate that the world's largest technology firms could spend about $6,000 billion on AI infrastructure between 2026 and 2031. Investors continue to price in strong future growth, but those expectations rest on companies' ability to convert today's record investments into profitable products and services. In many cases, evidence of such monetisation is not yet clear, raising questions about long-term returns.
Open-source models increase pressure on proprietary systems
The rapid improvement of open-source AI models is reshaping the market. As these models deliver better performance, they offer companies competitive, significantly lower-cost alternatives to premium, closed systems. This dynamic could fundamentally change competitive behaviour and incentivise firms to focus less on achieving the highest possible model performance and more on generating tangible business value.
What this means for companies
Szitás Lóránt, market analyst at XTB, emphasised: “The AI race is shifting from technological supremacy to economic sustainability. In the long run, winners may not be the firms that build the most powerful models, but those that can create tangible business value at sustainable costs.”
Based on the analysis, companies that integrate AI into products and services in a cost-conscious, effective manner—and that can deliver the growth and profitability investors expect—are likely to be better positioned in the coming years.
