Issue 01 — March 2026
IT EN DE
The European magazine on private AI
Market

AI market Italy 2025: 1.8 billion euros, but SMEs are falling behind. Where the untapped potential lies

Data from Politecnico di Milano's AI Observatory and Anthropic's research reveal a huge gap between AI capabilities and actual business usage. The numbers and what they mean for European SMEs.

AI marketPolitecnico di MilanoAI adoptionSMEstatisticsAnthropic

The Italian AI market: 2025 numbers

The Artificial Intelligence Observatory at Politecnico di Milano released updated data in February 2026. The numbers confirm strong growth — but also a widening gap.

1.8 billion euros: the value of the Italian AI market in 2025, growing +50% versus 2024. The 3-year CAGR 2022-2025 is +54%. Generative AI is the main driver, but not the only one: 54% of the market still comes from traditional Machine Learning projects.

The most relevant finding for businesses: 77% of the market involves custom projects, built bespoke for individual clients. Scalable models (SaaS, software licences) are growing faster, but ad-hoc projects remain the norm — confirming that effective AI is AI built on a company’s specific needs.

The gap between large and small companies

Here’s the data that should concern every European business leader:

AI adoption in Italian companies (2025) Source: AI Observatory, Politecnico di Milano — 698 companies surveyed Large enterprises 71% 10% 13% 6% Medium companies 15% 34% 33% 18% Small companies 7% 18% 42% 33% Active projects Early evaluation Interest but no initiative No interest

The numbers speak clearly:

  • 71% of large enterprises have active AI projects — up from 59% in 2024
  • 15% of medium companies have active projects
  • Only 7% of small companies have active projects
  • 75% of small companies haven’t started anything yet

47% of workers already use AI (but don’t tell IT)

Another finding: 47% of Italian workers say they use AI tools at work. But only 19% use exclusively company-provided tools. The rest use ChatGPT or Claude with personal accounts — Shadow AI. Company data ending up on third-party servers with zero oversight.

The untapped potential: the chart every CEO should see

Anthropic’s research offers a global perspective on the potential still to capture. The chart below compares, for each occupational category, the theoretical AI coverage with actual observed usage:

Theoretical AI capability vs actual usage by sector Source: Anthropic — The Anthropic Model Spec (2025) 0% 25% 50% 75% 100% Legal Finance IT / Software Engineering Education Media / Creative Healthcare Management Sales Administration Production Theoretical AI capability Actual observed usage

The message is clear: in nearly every sector, current AI can already cover 40%–90% of work activities, but actual usage stops at 15%–35%.

The widest gaps are in knowledge-intensive sectors: legal, finance, engineering, IT. Exactly the sectors where European SMEs operate and compete.

This gap isn’t technological — the technology exists. It’s an adoption, implementation and access gap. Large companies are closing it with internal teams. SMEs need a different path: ready solutions, accessible, that don’t require building AI capabilities from scratch.

What this means for your business

Solutions like ORCA by HT-X are designed to bridge this gap: private AI, on-premise or EU cloud, accessible to SMEs without internal AI teams. You don’t need to build capabilities from scratch — you need a partner that already has them.


Data: Osservatorio Artificial Intelligence, Politecnico di Milano (February 2026); Anthropic, “The Anthropic Model Spec” (2025). Charts are our own elaborations based on cited data.

Frequently asked questions

According to the AI Observatory at Politecnico di Milano, 71% of large enterprises have at least one active AI project. But the picture changes dramatically for SMEs: only 15% of medium-sized companies and 7% of small ones have active projects. 42% of small companies have no initiative and show no interest.

In 2025, the Italian AI market is worth 1.8 billion euros, growing 50% versus 2024. The 3-year CAGR 2022-2025 is 54%. Text analysis and conversation systems represent the largest share (39%), followed by data exploration and prediction (30%).

The main barriers are: difficulty estimating cost-benefit ratio upfront, lack of internal skills, unorganised data, and the perception that AI is 'for large companies only'. In reality, solutions like ORCA by HT-X are designed specifically for SMEs: accessible fixed costs, no technical expertise required, and deployment in weeks.

Looking for a private ChatGPT for your business?

ORCA is the on-premise AI platform by HT-X (Human Technology eXcellence): your data stays yours, GDPR and AI Act compliant.

Discover ORCA