The study's thesis
QuantumBlack, the McKinsey unit dedicated to artificial intelligence, published the study The AI value curve on October 14, 2026, which, according to the source, calculates, task by task, where AI costs less than the work it supports or replaces with equal result quality. The global opportunity is estimated at 2,520 billion dollars a year.
The authors' premise is that many organizations continue to measure adoption instead of value, counting users and tokens consumed: indicators that show how much AI is used but not how much it yields. The study therefore shifts attention from usage volume to the economic comparison between the cost of the technology and that of the work it acts on.
Why the calculation changes in Italy
The relevant point for decision-makers in Italian companies, according to the source, is that AI value depends on local labor cost. Consequently, the cost-effectiveness threshold is not universal: it shifts according to wages in the country where the technology is adopted.
This leads to a methodological warning: a business case built on US benchmarks tends to overestimate returns for Italian companies. The reasoning holds to the extent that US wages are higher than Italian ones, a circumstance the material assumes implicitly but does not quantify. Indeed, the source lacks any data on the pay gap between the two markets.
What remains to be verified
The material does not disclose the tasks analyzed, the sectors covered, the reference wages used for Italy, or the method by which the global figure of 2,520 billion was calculated. It is therefore impossible to establish which activities are cost-effective in the Italian context and which are not, or by how much business cases based on US parameters actually overestimate returns.
The source is also an article by Fabio Lalli, an innovation and AI consultant with more than twenty-five years of experience, founder of Iconico and of ZeroFive.ai, a consulting firm focused on AI adoption. It is therefore an expert commentary interpreting the study, not original research: assessments of the Italian case should be attributed to the author and verified against QuantumBlack's full document before being used as a basis for investment decisions.
