Brains on sale: how Italy trains AI talents and gives them away to Berlin and London

A hard suitcase and a badge from the Politecnico di Milano dangle from a student's backpack. Destination: Berlin. On the phone, an offer of 55-60 thousand euros a year for a junior role in artificial intelligence engineering. The last proposal in Italy? Only 33 thousand euros. The last call for London? 52-58 thousand pounds. The gap is not a detail: it is the price of the social elevator in the economy of generative models. And it explains, better than any slogan, why even in the “Year 2026” Italy continues to “train and lose” strategic skills.

The gap that bites: up to 40-50% between Italy, Germany, and the United Kingdom

Italy: a junior profile in AI/ML typically starts from 30-35 thousand euros gross, with independent estimates placing the typical value between 31 and 40 thousand depending on area and sector.

Germany: for the same role, the entry range often stands at 55-70 thousand euros gross.

United Kingdom: the most recent entry-level figures are around 50-55 thousand pounds (about 58-64 thousand euros at the average 2025 exchange rate), with scales widening rapidly where skills are scarce.

The result is a stable differential of around 40-50% against Italy in the entry-level AI engineering sector. This leads to a “magnet effect” that feeds on itself: those who leave earn more, accumulate experience in ecosystems with deep demand, and if they return, they often do so under conditions that are not always competitive. A snapshot that coincides with findings from the national press: a country “raising” AI talents that are then valued elsewhere.

What’s under the hood of the gap: productivity, taxes, demand

Labor productivity. The gap in values per hour worked remains wide compared to European peers: Istat series from 1995-2024 indicate weak dynamics of GDP per hour worked and a reduced contribution from total factor productivity. The 2024 updates indicate a decline in hourly productivity (-1.4%), a symptom of a fragile competitive base that also limits wage margins in knowledge-intensive sectors.

Tax wedge and labor cost. In 2024, the tax wedge for an average single worker rose to around 47%, among the highest in the OECD; the “Taxing Wages 2025” report places Italy in the top 10 for labor costs, compressing net pay without additional efforts from companies.

Internal question. While accelerating, Italian companies' spending on AI remains concentrated on a few large players and a limited number of use cases. According to the Artificial Intelligence Observatory of the Politecnico di Milano, the national market has risen to 1.2 billion euros in 2024 (+58% year-on-year) and about 1.8 billion in 2025 (+50%), but penetration among SMEs remains low compared to European potential. At the EU level, the share of companies with at least 10 employees using AI technologies reached 20% in 2025 (up from 13.5% in 2024), indicating that Italy is growing but not leading.

In summary, when productivity and demand base do not push sufficiently, AI risks remaining a “technology upgrade” without adequate salary compensation, especially in the early career stages.

An ecosystem that runs, but not enough

Increasing adoption, driven by the large

By 2025, 20% of European companies will use at least one artificial intelligence technology, a significant increase from 8% in 2023 and 13.5% in 2024. The growth is robust but uneven: large companies have shifted into high gear (with shares over 50% in various countries), while SMEs lag behind, slowing the spread of well-paid roles.

In Italy, projects involving generative AI are also concentrated in larger groups, focusing on text automation, demand forecasting, anomaly detection, and customer operations; upskilling along smaller supply chains is progressing in fits and starts.

The market generates volume, not always value

The expansion to 1.8 billion euros in 2025 certifies that demand exists; however, to translate this into competitive salaries, it requires: more end-to-end initiatives with measurable impacts on revenues, costs, and margins; greater data integration and MLOps to scale use cases beyond the pilot phase; a talent pool with “T-shaped” skills (methods, coding, domain) to reduce time-to-value.

Salaries: where the price is formed (and how to raise it)

The salary range for a junior AI/ML professional in Italy is influenced by five determinants: local availability of scalable projects and not just “proof of concept”; competition among employers for scarce skills (including global players); quality of the overall package (salary, bonuses, stock options, training, remote work); net labor cost after contributions and taxes; expectations for professional growth, meaning the credible promise of moving from 35k to 50-60k in 2-3 years.

The Germany benefits from strong vertical markets (industry, automotive, advanced manufacturing) that adopt AI for efficiency and new offerings, supporting entry ranges from 55 to 70 thousand euros. The United Kingdom, with a developed fintech ecosystem and a long data-driven tradition, positions the initial steps above 50 thousand pounds. Italy, while closing part of the digital gap, suffers from the "bottleneck" between prototype and production and a tax burden that affects the net pay.