AI in Business: Strong Adoption, Limited Impact on EBIT

Companies are deploying more AI and allocating increasing budgets to it, but the impact on profits remains modest. Global data from McKinsey and French surveys from INSEE, Eurostat, and the ECB paint a consistent picture: high adoption, still limited profitability, and France slightly lagging in investment efforts.

In France, 7.2% of planned investments allocated to AI, compared to 9.1% in the euro area

In the fourth quarter of 2025, the ECB surveyed 4,968 companies, including 625 French ones, about their investment plans. French companies plan to allocate 7.2% of their total investment to AI, while the average in the euro area reaches 9.1%. This gap of 1.9 points is stable and reflects the observed adoption differential. In July 2026, INSEE reported that 18% of companies in France were using at least one AI technology, which is 8 points higher than in 2024. Eurostat places the European average at 20%, with France two points below, after an increase from 13.5% twelve months earlier. Adoption varies by size and sector: 58% of companies with 250 employees or more use AI, as do 59% in information and communication. Among non-user French companies, 71% see no utility in it, and 54% cite a lack of expertise. The organizations clarify that these surveys are based on distinct scopes and do not accumulate, but they outline a coherent set: adoption close to the European norm, slightly lower financial effort, and profitability yet to be demonstrated for the majority.

Impact on EBIT: 37% report an effect, most often below 5%

According to McKinsey, 37% of respondents attribute at least some impact on EBIT to their use of AI. The 2025 edition of the barometer recorded 39%, and the two-point decline is presented as remaining within the margin of variation of a declarative survey. Within this group, the majority report the effect as being below 5% of EBIT. McKinsey estimates that AI is "on the road to return on investment" and notes a two-point differential compared to the previous edition.

Broad adoption, full deployment rare in organizations

In 2025, 88% of respondents reported regularly using AI in at least one function, but only 7% indicated deployment at the organizational level. The gap between usage and financial contribution remains the central issue. Two years earlier, Gartner anticipated that one-third of AI projects would be abandoned.

Method, expectations, and scope of McKinsey

The barometer The State of AI 2026 is based on 1,719 respondents worldwide and is part of McKinsey's annual publications on AI in business. In December 2025, experts were still projecting 2026 for the profitability of generative AI. McKinsey also markets AI transformation missions for large organizations, a context that surrounds its interpretation of the results.