AdoptionAbandonment

Companies not using AI: who is holding back, and on what grounds

Discover which companies are not using AI and the official 2025 data behind non-adoption. Legal uncertainty and compliance remain the primary barriers.

, AI Researcher5 min read
In this article7
  1. The named companies actually holding back
  2. What each has stated, and where
  3. The reasons given: risk, cost, evidence
  4. What silence does not prove
  5. What non-adoption tells a planner
  6. Frequently asked questions
  7. Sources
Abstract representation of corporate compliance and data security, featuring a European enterprise building shielded by regulatory motifs, conveying the concept of companies holding back on AI.
Abstract representation of corporate compliance and data security, featuring a European enterprise building shielded by regulatory motifs, conveying the concept of companies holding back on AI.

The named companies actually holding back

The answer depends on which measurement is read. Statutory offices survey enterprises with defined employee thresholds, while industry associations survey their own panels. Eurostat, the EU statistical office, surveyed enterprises with 10 or more employees in early 2025 and found that 19.95% used at least one AI technology (retrieved 23 September 2026)[1]. Destatis contributes the German figures to this same EU survey collection (retrieved 23 September 2026)[2]. Many enterprises face strict regulatory requirements; checking whether an AI practice is permitted under the AI Act involves reviewing its prohibited practices. Furthermore, companies must consider data residency when determining if AI is available in Europe, leading some to review their abandoned AI pilots.

The table below states, for each named publisher, what it publishes and on what basis. The distinction matters because a consultancy survey of organisations worldwide is not a statistical-office count of enterprises in one jurisdiction: the recurring error on this topic is reading a vendor survey as a statistical-office figure. An en dash in the table means the provider's page lists no value for that column.

PublisherType of bodySurvey populationLatest AI use figureBarriers measured
BitkomGerman industry association603 German companies with 20 or more employees (Bitkom Research)57% used AI in 2026; 4% said AI was not a topic for them, down from 41% two years earlier (retrieved 23 September 2026)85% of non-users cite missing technical know-how, 66% legal uncertainty, 63% hard-to-calculate costs (retrieved 23 September 2026)
DestatisGerman federal statistical officeGerman enterprises with 10 or more employees, as the German contribution to the EU ICT usage surveyGerman AI use figure published via the ICT survey and GENESIS database (retrieved 23 September 2026)–
European Investment BankEU bank, investment survey unitApproximately 13,000 EU firms plus a US sample––
EurostatEU statistical office157,000 of 1.53 million EU enterprises with 10 or more employees, surveyed in early 202519.95% used at least one AI technology in 2025; 55.03% of large enterprises (retrieved 23 September 2026)70.89% lack of expertise, 52.52% lack of clarity about legal consequences, 48.83% data protection and privacy concerns (retrieved 23 September 2026)
INSEEFrench national statistical instituteFrench enterprises with 10 or more employees––
IstatItalian national statistical instituteItalian enterprises with at least 10 employees––
McKinsey & CompanyManagement consultancyOrganisations worldwide (vendor survey, not a statutory count)––
OECDIntergovernmental economic organisationCross-country firm-level AI adoption statistics and studies––
Office for National StatisticsUK national statistical officeUK businesses via the Business Insights and Conditions Survey––
Stanford HAIAcademic research institute (Stanford University)Annual AI Index collating global data across countries and sectors88% organizational adoption reported for 2025 (retrieved 23 September 2026)Responsible AI reporting gaps; 362 documented AI incidents in 2025, up from 233 in 2024 (retrieved 23 September 2026)

What each has stated, and where

Against these figures, the Stanford HAI 2026 AI Index reports organizational adoption reaching 88% (retrieved 23 September 2026)[4], a figure based on its collated organizational surveys rather than a statistical office's enterprise count. The two readings are not contradictory: they measure different populations. Statutory surveys count all enterprises above a size threshold, including the small firms that dominate the population, while organizational surveys skew towards larger organisations. Non-adoption figures here capture measured real-world use across different enterprise sizes, not general market sentiment.

The reasons given: risk, cost, evidence

Germany's industry association Bitkom measures the same pressures in its 2026 survey: among companies without AI, 85% cite missing technical know-how for implementation, 66% cite legal hurdles and uncertainty, and 63% cite hard-to-calculate costs (retrieved 23 September 2026)[3]. On the data protection side, 9 out of 10 companies demand reform of data protection supervision, and in every second company data protection hinders the use of AI (retrieved 23 September 2026)[3]. Regulatory exposure is concrete: the EU AI Act is now in force, leading many enterprises to closely evaluate their compliance obligations (retrieved 23 September 2026)[5].

For a reader checking what a rule requires, two facts matter for interpreting these figures. First, the AI Act has applied since 2 February 2025 to its prohibited practices, with further obligations phased in through 2028; the Act regulates AI systems and their providers and deployers, and does not prohibit enterprise AI use as such (retrieved 23 September 2026)[5]. Second, Bitkom's 57% AI use among German companies in 2026[3] and Eurostat's 19.95% EU enterprise figure[1] describe different populations, questions and reference years, and both are traceable to the pages cited above.

What silence does not prove

Official reports indicate that depth of integration is often a better measure of holding back than outright non-use. While many enterprises report some level of AI activity, the scope is frequently limited to a single business process rather than widespread deployment.

What non-adoption tells a planner

Stanford HAI’s 2026 AI Index documents 362 AI incidents in 2025, up from 233 in 2024, showing that compliance gaps and responsible AI issues accompany the adoption rates (retrieved 23 September 2026)[4].

Frequently asked questions

What companies are rejecting AI?

Regulated enterprises and those without clear legal frameworks are actively delaying AI deployment. According to Bitkom in 2026, data protection requirements force many to hold back, with 9 out of 10 surveyed companies demanding reform of data protection supervision (retrieved 23 September 2026).

What companies are not affected by AI?

Official 2025 statistics from Eurostat show that AI adoption varies by economic activity; for example, sectors like construction report the lowest AI use at 10.79% (retrieved 23 September 2026). Companies in these sectors often find the technology less relevant to their immediate operational workflows.

What brands are against AI?

Rather than a public stance against AI, major European brands delay adoption due to strict compliance mandates. The EU AI Act imposes significant obligations, leading procurement teams to hold back until legal clarity is established (retrieved 23 September 2026).

Sources

  1. ec.europa.eu
  2. destatis.de
  3. bitkom.org
  4. hai.stanford.edu
  5. artificialintelligenceact.eu
  6. bitkom.org

Rankings and scores on Optinest cannot be bought — no commercial relationship changes how a provider is assessed. How we score

Related reading

All Adoption →

What industries use AI the most

Discover what industries use AI the most based on 2026 official statistics. Compare AI adoption rates across sectors, enterprise sizes, and data publishers.

A minimalist structural diagram illustrating varying levels of enterprise AI adoption across different industry sectors and company sizes.