The state of AI in the Dutch MKB 2026: adoption statistics for the Netherlands

The state of AI in the Dutch MKB 2026: the headline numbers

If you run a Dutch SME and feel behind on AI, the official data says you are in good company — and that the gap is closing fast. According to CBS, Statistics Netherlands (2024), 22.7% of Dutch companies with ten or more employees used at least one of seven listed AI technologies in 2024 (rounded to 23% in CBS summaries). That is a sharp jump: CBS reports the figure rose from 14% in 2023 to 22.7% in 2024, an increase of nearly nine percentage points in a single year.

What are companies actually doing with it? The most-used technology is text mining — the analysis of written text — at 13.5%, around two-and-a-half times its 2023 level, per CBS (2024). Close behind is natural language generation (producing written or spoken text) at 12.3%, roughly three times the 2023 figure. By business purpose, CBS reports that among AI-using companies the most common application is marketing and sales (36%), followed by administrative processes and management tasks (30%), per the CBS Dutch AI Monitor 2024 summary. In other words, the early wins are in language-heavy, repetitive work — exactly the tasks that fill an SME's week.

One number reframes the whole picture. CBS (2024) reports that AI-using companies already account for over half (51%) of total Dutch company revenue. Adoption is concentrated among larger, higher-revenue enterprises — which means the practical question for the MKB is not whether AI works, but how to apply it without an enterprise budget.

Adoption by company size and sector — and how the Netherlands compares to the EU

The size gradient is the most important pattern in the data for any SME owner. According to the CBS Dutch AI Monitor 2024, AI use climbs steadily with headcount: 17.8% of companies with 10–19 employees, 22.2% at 20–49, 28.1% at 50–99, and 34.8% at 100–249. At the top end, CBS (2024) reports 59.2% of companies with 500 or more employees use AI. The smallest measured bracket sits at less than a third of the largest — a real adoption gap between the MKB and big enterprise, but also a clear sign of headroom.

Which sectors lead

By sector, information and communication (ICT) leads decisively. CBS (2024) reports 58.0% of ICT companies used AI in 2024, up from 37.0% in 2023. CBS, The Netherlands in Numbers 2025 puts specialised business services next at 40%, then financial services at 37%. If you are outside these front-runner sectors, the takeaway is not that AI does not apply to you — it is that adoption in your sector is still early enough to be a genuine advantage rather than table stakes.

The European comparison

The Netherlands sits comfortably above the European average. On the Eurostat-comparable basis (which excludes the financial sector), CBS (2024), citing Eurostat, puts Dutch adoption at 23.1% against an EU average of 13.5% — roughly 1.7 times the EU figure — ranking the Netherlands 6th in the EU-27, behind Denmark, Sweden, Belgium, Finland and Luxembourg. The European baseline is rising fast too: Eurostat (2025) reports that 20.0% of EU enterprises with ten or more employees now use AI, up 6.5 percentage points from 13.5% the year before. Eurostat (2024) shows the earlier step from 8.0% to 13.5% across 2023–2024. The frontier countries are pulling away: Eurostat (2025) reports Denmark at 42.0%, Finland at 37.8% and Sweden at 35.0%, while Romania trails the bloc at 5.2%.

The main barriers for the MKB

The numbers above describe the barrier without needing a survey to name it: AI use rises almost perfectly with company size. The pattern in CBS (2024) — from 17.8% in the smallest bracket to 59.2% in the largest — points squarely at the resources that scale with headcount: budget, in-house data and technical skills, and the time to run a project at all. A smaller firm rarely has a data team to lean on or a spare quarter to experiment, so the cost of learning on a live, revenue-critical process feels higher.

The concentration figure reinforces the same point. With AI-using companies accounting for over half of total Dutch company revenue (CBS, 2024), the firms furthest ahead are the ones that could most easily absorb a failed experiment. For the MKB the constraint is the reverse: limited room for a project that does not pay back. That is an argument for a bounded, measured first step rather than a broad transformation — proving value on one process before committing budget to the next. It is also why regulatory clarity matters so much to smaller firms, which brings us to the rules that take effect this year.

The EU AI Act 2026 timeline you need to plan around

The EU AI Act is the regulatory backdrop to every Dutch AI project from here on, and 2026 is a pivotal year in its rollout. According to the European Commission and the official AI Act implementation timeline, the key dates are:

• 1 August 2024 — entry into force. The Act became law, but its obligations were phased in rather than applying immediately.
• 2 February 2025 — prohibited practices and AI literacy. Bans on certain unacceptable-risk AI uses, plus AI-literacy obligations for staff, began to apply.
• 2 August 2025 — general-purpose AI (GPAI) models. Rules for providers of general-purpose AI models became applicable.
• 2 August 2026 — the main milestone for high-risk systems. Most high-risk obligations begin to apply, marking the point at which the Act becomes broadly applicable.
• 2 August 2027 — remaining high-risk rules. The final tranche of high-risk obligations, including those tied to products covered by existing EU safety legislation, takes effect.

For most MKB use cases — drafting marketing copy, summarising documents, internal search over your own files — you are unlikely to be operating a high-risk system. But the AI-literacy duty already applies, and the safest assumption is to design for compliance from the first proof of concept rather than retrofitting it later. Knowing which tier your use case falls into is itself part of the work.

What this means for the MKB — the Crux take

Read together, the data tells a consistent story. Adoption is real and accelerating, the early wins are in language and admin work, and the only thing separating the MKB from larger firms is resources — not relevance. The honest implication is that a smaller company does not need to match enterprise scale to benefit; it needs to pick the right single use case and prove it cheaply before scaling.

That is how we work at Crux Digits, a boutique AI consulting and implementation company in Nieuwegein, in the province of Utrecht, founded in 2022 and working bilingually in English and Dutch. We are an AI consultant for the MKB, and our delivery is deliberately staged to match the resource constraints the CBS numbers expose. It starts with a fixed-price AI Audit and Strategy at €2,500 (excl. VAT) to find the one process worth automating, a Proof of Concept at €20,000 to prove it on your real data, and a production launch from €50,000 once the value is demonstrated, with ad-hoc work at about €150 per hour outside that ladder. You can see the full breakdown on our pricing page. Every engagement is built EU AI Act- and GDPR-first, so the compliance milestones above are designed in, not bolted on.

We have delivered 13 case studies on this model, and our view is plain: the firms that win the next two years are not the ones that spend the most, but the ones that start with a bounded, measurable project. If you are weighing who to work with, our guide to the best AI consultants in the Netherlands sets out what to look for so you can judge any partner — including us — on the same terms.

Methodology and sources

This page reports only figures attributable to named, official statistical sources, each linked inline at the point it is used. The CBS figures come from the CBS Dutch AI Monitor 2024 and the related news release "Increasing use of AI by business", with sector and EU-comparison context from CBS, The Netherlands in Numbers 2025. European figures are from Eurostat (2025) and Eurostat (2024). Regulatory dates are taken from the European Commission and the AI Act implementation timeline. Where CBS reports both a national figure (22.7%) and a Eurostat-comparable figure (23.1%), both are shown, since they use slightly different methodologies. Percentages are reported as published; minor rounding appears in CBS lay summaries.

These public datasets describe the market in aggregate. Crux Digits will extend this picture with its own MKB AI survey — a proprietary layer focused specifically on smaller Dutch firms — to add the on-the-ground detail that national statistics, built around the 10-or-more-employee threshold, tend to miss.