18%
NL industrial companies using min. 1 AI application
ING Research, 2026
AI in Dutch manufacturing
Figures, ROI and trends. 50+ validated statistics on the current state of AI in industry. Updated May 2026.
%
of NL industrial companies use at least one AI application
ING Research, 2026
%
of AI pilots worldwide produce no measurable ROI
MIT Project NANDA, 2025
%
defect reduction with AI quality control in production
McKinsey, 2025
%
of large NL manufacturers (250+) use AI
ING Research, 2026
1. AI adoption in Dutch industry
Dutch industry stands at a tipping point. According to ING Research 18% of industrial companies now use at least one AI application. A doubling in two years. But behind that growth number lies wide variation, from nearly 14% at the smallest companies to 66% at large ones.
29%
NL manufacturers (10+ employees) using AI
ING/Eurostat, 2026
64%
Large NL manufacturers (250+) using AI
ING Research, 2026
AI usage by company size (2025)
Micro (2-10 emp.)
13.8%
Small SMB (10-50 emp.)
18%
Mid-size SMB (50-250 emp.)
45%
Large (250+)
66.2%
NL industry total (10+)
29%
Source: CBS ICT-use enterprises 2025, ING Research 2026
Strongest growth sits with mid-size companies. AI use at companies with 50 to 250 employees rose from 20% in 2023 to 45% in 2025. Exactly the segment many Dutch manufacturers occupy.
Source: CBS ICT-use enterprises (2025), ING Research thematic study "AI in industry" (February 2026), Eurostat Digital Economy Index 2024.
2. Netherlands vs Europe
Dutch manufacturers deploy AI more often than the EU average, but lag the European front-runners. The adoption gap with Belgium and Denmark is over 10 percentage points. Meanwhile software investments in Dutch industry grow much slower than the rest of the economy over the past five years.
#6
Netherlands position in EU on total AI business adoption
Eurostat, 2024
-7.5%
real value of NL industrial software capital shrunk (2019-2024)
ING Research, 2026
16 pp
gap with national growth of 8.5% software capital
ING Research, 2026
AI adoption manufacturing: Netherlands vs Europe (2025, 10+ emp.)
Belgium
40%
Denmark
40%
Netherlands
29%
EU average
17%
Source: ING Research / Eurostat, February 2026
Industrial companies have much to gain from successful AI deployment. They are well advised to take a pragmatic approach. Start small and bring in outside expertise where needed.
Source: ING Research thematic study "AI in industry" (February 2026), Eurostat Digital Economy & Society Index 2024, CBS ICT-use enterprises 2025.
3. Why 80 to 95% of AI pilots fail
The gap between "doing an AI experiment" and "running AI in production". Where we do our work.
The most striking figure in this overview: between 80 and 95 percent of all AI pilots worldwide produce no measurable ROI. MIT, BCG and RAND independently reach the same conclusion. 30 to 40 billion dollars invested worldwide, without return.
95%
of GenAI pilots produce no measurable ROI
MIT Project NANDA, July 2025
74%
of companies cannot scale AI
BCG, Oct 2024
80.3%
of enterprise AI projects fail entirely
RAND Corporation, 2025
28%
of AI use cases meet ROI expectations
Gartner, Apr 2026
33%
of companies achieved enterprise-wide AI
McKinsey, 2025
What happens to enterprise AI projects (RAND, 2025)
The problem is almost never the AI model. It is data readiness, missing workflow integration, and the absence of a predefined success criterion. That is exactly what we do differently. We do not pilot. We build straight to production. On your own infrastructure. With measurable ROI from month one.
Source: MIT Project NANDA "The GenAI Divide" (July 2025), BCG "AI Adoption in 2024", RAND Corporation "Enterprise AI Failure Research" (2025), Gartner Survey 782 I&O Leaders (April 2026), McKinsey State of AI 2025.
4. What AI in manufacturing does deliver
ROI figures per application, validated from public research.
Alongside those alarming failure rates we see what AI in production environments does deliver when implemented well. Four main applications: quality control, predictive maintenance, supply chain optimisation, and general productivity.
4.1 Quality control
AI machine vision detects defects in milliseconds with 99%+ accuracy. Traditional manual inspection misses between 20% and 30% of defects. The gap is no longer incremental, it is a competitive divide.
30-50%
reduction in defect rates with AI machine vision
McKinsey, 2025
99%+
AI machine vision detection accuracy
Applied AI Studio, 2026
200-300%
ROI on full AI quality control implementation
Tech-stack analysis, 2026
6-12 mo
average payback time quality control
Tech-stack analysis, 2026
4.2 Predictive maintenance
Predictive maintenance uses sensors, IoT data and AI algorithms to predict machine failures before they occur. The difference with traditional maintenance: less downtime, lower costs, longer equipment life.
25%
improvement in Overall Equipment Effectiveness (OEE)
Case study processed food manufacturer
30%
lower maintenance costs through AI-driven planning
SmartDev, Dec 2024
25%
reduction in equipment downtime via real-time monitoring
Business Research Insights
90%
reduction in assembly defects with AI assembly verification
Applied AI Studio, 2026
4.3 Supply chain and inventory
AI improves inventory management and supply chain decisions by detecting real-time patterns human planners miss.
150-250%
ROI on AI supply chain optimisation
Tech-stack analysis, 2026
200%
more supply chain disruptions detected via AI quality monitoring
AllAboutAI, 2025
4.4 General productivity
Independently of specific application, AI implementation drives structural productivity growth. Two independent studies reach comparable conclusions.
3 pp
extra annual employee productivity growth with AI implementation
ING Research based on academic studies
4%
increase in labour productivity from AI adoption
BIS/EIB study on 12,000+ EU companies
Source: McKinsey research on AI quality systems, BIS Working Paper "AI adoption, productivity and employment" (BIS/EIB, 2025), Applied AI Studio "AI Use Cases Transforming Manufacturing Quality Control", Tech-stack "AI Adoption in Manufacturing: ROI Benchmarks", Sandia National Laboratories research.
5. Market growth and investment
The industrial AI market is growing fast. Worldwide the AI in manufacturing market grows roughly sevenfold towards 2030. The European predictive maintenance market grows 27.5% per year. A window: early adopters build a lead that cannot be closed in a few years.
$17.44B → $115.76B
AI in manufacturing worldwide (2025 → 2030)
Applied AI Studio, 2026
$14.29B → $98.16B
Predictive maintenance worldwide (2025 → 2033), 27.9% CAGR
Grand View Research
$3.65B → $32.47B
Predictive maintenance Europe (2024 → 2033), 27.5% CAGR
Market Data Forecast
22.5%
Germany market share in European predictive maintenance market (#1)
Market Data Forecast, 2024
Source: Grand View Research "Predictive Maintenance Market", Market Data Forecast "Europe Predictive Maintenance Market", Applied AI Studio AI in Manufacturing Market Report, Fortune Business Insights, Mak Data Insights (February 2026).
6. Dutch CEOs and sentiment
Dutch leadership is aware of what is happening. 90% of top executives accelerate AI applications in 2026. At the same time AI use tops the worry list. Optimism and concern walk hand in hand.
90%
NL top executives accelerate AI applications in 2026
ING CEO Survey, Dec 2025
92%
is (very) optimistic about 2026 (was 85% in 2025)
ING CEO Survey, Dec 2025
25%
expected change in current job roles
ING CEO Survey, Dec 2025
#1
AI use tops the CEO worry list
ING CEO Survey, Dec 2025
Rapid AI developments will increasingly influence the operations and investments of Dutch companies in 2026.
Companies that considered AI but ultimately do not use it cite one main reason: lack of experience. 74.6% according to CBS. That is exactly what external expertise solves.
Source: ING CEO Survey December 2025, CBS "Use of artificial intelligence by companies increases" (September 2025).
7. AI tools in business
Which AI models and tools dominate enterprise use in 2026?
An overview of the most deployed platforms, based on enterprise adoption and public market data. The gain is rarely in one tool, but in the combination.
In practice no modern company uses one single AI tool. The gain is in the combination. ChatGPT or Claude for analysis, Microsoft Copilot for M365 integration, GitHub Copilot for code, n8n for workflow automation. We use Claude daily for strategic copy and code review, n8n for agent orchestration, and custom-built AI systems for production deployment at clients.
Source: IntuitionLabs "Claude vs ChatGPT vs Copilot vs Gemini: 2026 Enterprise Guide", DataNorth AI "Top 10 Best AI Tools for 2026 Q2 Update", Searchlab "Best AI Tools 2026".
8. Future predictions
What do AI leaders themselves expect? The most striking predictions, ordered by year.
Important: these are predictions, not facts. Nobody can predict the future. We share them to sketch the direction, not as certainty.
2026
AI will surpass the smartest human intelligence within 1 year
Elon Musk (xAI), public statements, 2025
2027
AI will surpass humans in nearly everything
Dario Amodei (Anthropic), public statements, 2025
2027-2030
$13 trillion in additional global economic activity from AI by 2030
McKinsey Global Institute
2030
AI will surpass human intelligence
Sam Altman (OpenAI), interview Die Welt 2025
2030
30 to 40% of current economic tasks will be performed by AI
Sam Altman (OpenAI), Fortune interview 2025
2030
1.2% extra global GDP growth per year from AI
McKinsey "Notes from the AI frontier"
Robotisation is not a luxury but a precondition to preserve our manufacturing industry. A national agenda gives direction, accelerates adoption, and ensures companies of all sizes get access to technology that shapes their future.
The above are predictions by CEOs and research institutes. Actual developments may differ. See also our methodology and disclaimer at the bottom of this page.
Source: Fortune "Sam Altman thinks AI will surpass human intelligence by 2030", Die Welt interview Sam Altman 2025, McKinsey "The economic potential of generative AI", McKinsey "Notes from the AI frontier", TNO publication Smart Industry 2026, Industrie Magazine April 2026.
Key conclusions
Six insights every industrial decision-maker should account for.
1
The Netherlands is lagging
NL manufacturing sits at 29% AI adoption while Belgium and Denmark are already around 40%. Software investments have shrunk for five years. The gap is widening, the pace is not.
2
Pilots are not the answer
80 to 95% of AI pilots worldwide produce no measurable ROI. Companies that do scale opt for direct production deployment. No more experiments.
3
ROI is proven when done right
AI quality control delivers 30 to 50% defect reduction. Predictive maintenance improves OEE by 25%. Quality control ROI: 200 to 300% within 6 to 12 months.
4
Manufacturing is a growth market
The European predictive maintenance market grows from $3.65B (2024) to $32.47B (2033). AI in manufacturing worldwide: from $17.44B to $115.76B towards 2030.
5
Dutch leadership knows it
90% of Dutch top executives accelerate AI applications in 2026. Main reason not to: lack of experience (74.6%). That is exactly what external expertise solves.
6
The future moves fast
AI leaders predict AI will surpass human intelligence between 2027 and 2030. McKinsey estimates $13 trillion extra in world economy by 2030. Waiting has a price.
Methodology and sources
The statistics on this page are compiled from publicly available research reports by reputable organisations. We refresh this page regularly with the latest data.
Last updated: May 2026
- ING Research
- Eurostat
- MIT Project NANDA
- BCG Boston Consulting Group
- RAND Corporation
- Gartner
- McKinsey Global Institute
- BIS / EIB
- Grand View Research
- Market Data Forecast
- Applied AI Studio
- Smart Industry NL, FME, AIC4NL
- TNO
Disclaimer
Figures on this page are drawn from the most recent available editions of the cited reports. Some statistics concern preliminary results or estimates.
Future predictions are opinions of CEOs and research institutes. They are not facts and not guarantees. Nobody can predict the future.
This page is compiled for information and is not financial, strategic or legal advice. VoidTech Solutions accepts no liability for decisions made based on the information on this page.
For specific advice contact us via patrick@voidtechsolutions.com.
VoidTech perspectief
What do we do with this?
A statistics page is interesting to read. But the figures above, especially those on pilots-that-fail, are exactly why we started VoidTech.
We do not pilot. We build working AI systems directly in your production environment. On your own infrastructure. With measurable ROI from month one. Not experimenting with data readiness, but getting the data in order first. Not talking about use cases, but building one that works.
Patrick Leegte worked many years in Dutch manufacturing before going full-time on building AI systems. He knows the difference between something that works in a demo and something that still works in production at three in the morning. That difference is the whole story.
Frequently asked questions
“
How many Dutch manufacturers use AI in 2026?
According to ING Research 18% of Dutch industrial companies use at least one AI application. Measured among companies with 10 or more employees the figure is 29%. At large manufacturers (250+ employees) it is 64%.
“
How does the Netherlands compare to other European countries?
The Netherlands ranks 6th in the EU for AI adoption. Front-runners are Denmark (27.6%), Sweden, Belgium, Finland and Luxembourg. In manufacturing specifically the Netherlands sits at 29% while Belgium and Denmark already score around 40%.
“
What is the average ROI on AI projects in industry?
ROI varies strongly by application. AI quality control achieves 200 to 300% ROI within 6 to 12 months. AI supply chain optimisation 150 to 250%. Predictive maintenance reduces maintenance costs by 30% on average.
“
Why do so many AI projects fail?
MIT Project NANDA (2025) concludes 95% of GenAI pilots produce no measurable ROI. The main causes are data readiness, missing workflow integration, and the absence of a predefined success criterion. The problem is almost never the AI model itself.
“
Which AI tools are most used by companies?
ChatGPT, Claude, Microsoft Copilot and Google Gemini are the most used enterprise AI platforms. For code: GitHub Copilot and Cursor. For workflow automation: n8n. For visual content: Midjourney. Most companies do not use one tool but a combination.
“
What is predictive maintenance exactly?
Predictive maintenance uses sensors, IoT data and AI algorithms to predict machine failures before they occur. Instead of fixed maintenance intervals (preventive) or reacting after failure (reactive), maintenance is planned based on the actual condition of the equipment.
“
How much should an SMB manufacturer invest in AI?
That depends strongly on the application. A first AI quality control system runs between €25,000 and €150,000 including integration. A first predictive maintenance implementation sits in the same range. Our rates are transparent: €110 to €125 per hour, or fixed-price after the scoping phase.
“
Which industry sectors apply AI most in the Netherlands?
According to CBS 2025: information and communication (54%), specialised business services (40%), financial services (37%). Manufacturing sits at 29%. Logistics lags most at 4% AI use.
“
What does the EU AI Act mean for production companies?
Since 2 February 2025 companies that develop or use AI are required to make staff AI literate. That means: enough knowledge of the AI systems they use. For high-risk AI applications stricter requirements apply around transparency and risk management.
“
How do I start with AI in my production company?
Start with one concrete problem you want to solve, not with the technology. Examples: speed up work preparation, service ticket processing, predictive maintenance on one machine, quality control on one product line. Start small, get quick results, then scale. Avoid building a 18-month roadmap.
👤
Written by
Patrick Leegte
Founder VoidTech Solutions, years of manufacturing experience, multiple proprietary platforms in production
Patrick has many years in Dutch manufacturing. Since 2024 he builds AI systems for industrial companies full-time. He previously worked at or for Shell, Wärtsilä, Lagersmit, IHC, FN Steel and Heavy Cargo Lifters. This page was compiled based on public research sources. Questions or comments on the content? Get in touch.
Schedule a free intro call
No sales pitch, no pilots. An honest conversation about what works in your situation.
Schedule an intro call