Are euro-area banks ready to judge AI loans? Most told the ECB they still lack specific metrics
In a 159-bank survey, most lenders said they had not introduced quantitative criteria for AI-related lending. Banks expect demand to rise gradually, especially from data-centre and infrastructure providers, but the survey did not measure actual AI loan volumes.
The short version
- What happened
- The ECB surveyed 159 euro-area banks between 21 September and 6 October and reports a 100% response rate.
- Who it affects
- Businesses building data centres may find lenders more prepared to discuss AI-related finance than smaller firms buying software, while inconsistent criteria could make applications harder to compare.
- What remains unproven
- Bank-reported survey
Keep in mind
Bank-reported survey
AI answers use qualitative counts
No AI lending-volume denominator
No pricing, approval or default outcomes
Expectations are not forecasts
Country-level AI responses are not shown
Source analysis · 1 linked source. Primary analysed source date: 27 October 2026.
Keep following the evidence
Join The Impact Brief founding list for selected AI reports, their effect on people and the evidence limits. Read a sample first.
Read the full report · 4 min The evidence, methods, implications and limitations
AI is entering credit conversations before a common scorecard exists
Most euro-area banks responding to a new ECB question said they had not yet added specific quantitative criteria for loans connected with developing or adopting artificial intelligence. A small number said AI adoption and its risks increasingly appear in qualitative borrower assessments, with information-technology companies receiving the closest scrutiny. That suggests judgement is entering conversations before a shared underwriting scorecard has emerged.
For a company seeking finance, this can make evidence quality unusually important. A credible application may need to separate a contracted data-centre asset from speculative capacity, disclose customer concentration and power commitments, and show how software or automation revenue will cover debt service. The ECB survey does not say which metrics banks will adopt, whether similar borrowers receive similar treatment or whether AI exposure changes the price of credit.[1]
Expected demand points toward infrastructure, not every AI adopter
Several banks expect a gradual but sustained increase in demand for loans financing digital infrastructure, particularly data centres. They said current and prospective AI-related lending is more likely to target infrastructure providers than software developers or firms adopting AI, and they expect that emphasis to strengthen over coming months. This is an expectation reported by lenders, not a measured forecast with an amount, growth rate or probability interval.
The wider survey shows a tougher credit backdrop. A net 9% of banks reported tighter standards for loans to firms in the third quarter, while firm loan demand fell on a net basis by 6%. Higher interest rates were the main drag. Those net percentages are balances between respondents reporting increases and decreases; they do not mean credit standards rose by 9% or that loan volumes fell by 6%.[1]
The distinction that matters
What the ECB survey measured
The release mixes observed lending conditions, bank expectations and a new qualitative AI question.
Observed
Banks reported third-quarter credit conditions
Net 9% tightened standards for firms
Expected
Several banks described future AI-related demand
Gradual growth, weighted toward infrastructure
Not measured
AI loan amounts, prices and defaults
No denominator or outcome series published
The denominator is strong for participation, weaker for the AI claim
The Eurosystem surveyed 159 banks from 21 September to 6 October 2026 and received a response from every bank. Results mainly describe changes observed in the third quarter and expectations for the fourth. That provides a clear institutional denominator, but the release uses terms such as 'most', 'a small number' and 'several' for the dedicated AI answers rather than publishing exact respondent counts or country splits.
A future assessment would be stronger with the AI questionnaire tables, actual lending amounts, pricing and rejection outcomes, borrower types and performance after origination. Until then, the defensible finding is narrower: banks see potential demand around physical infrastructure, yet most have not translated AI-specific risks into disclosed quantitative lending criteria.[1]
Watch the background, then check the evidence
Illustrative photo · not a video frame
UNESCO · 11 min 27 sec · Background explainer
UNESCO Talks: Interview with 2024 Nobel laureate Professor Daron Acemoglu
YouTube loads only when you choose; playback starts when you press play.
Economist Daron Acemoglu discusses how choices about AI affect workers, inequality and society. Pair the interview with our comparison of task exposure, hiring and skills evidence.
Evidence check: An expert interview from 2024, offering a perspective on policy and work rather than a measured forecast of job losses.
Published 2024-10-14. Source checked 14 October 2026. Check captions on the publisher’s video page.
Read the AI and jobs guidePhoto: Dylan Gillis / Unsplash · Licence
Continue the story
Read next
Continue exploring ai infrastructure: compare the finding and its evidence limits.
Why is Meta seeking up to $95bn while saying its AI investments already pay?
Meta launched about $35bn of share offerings and outlined a possible $60bn programme for 2027. The plan signals the scale of its infrastructure bill—not that all the money has been raised.
Read the next reportExplore the full source and evidence record
What this means for people
- Businesses building data centres may find lenders more prepared to discuss AI-related finance than smaller firms buying software, while inconsistent criteria could make applications harder to compare.
Global context
The survey covers the euro area and should not be generalized to banks operating under different capital, energy and data-centre markets. International lenders may use different definitions of AI-related credit.
What the evidence does not yet show
- Bank-reported survey
- AI answers use qualitative counts
- No AI lending-volume denominator
- No pricing, approval or default outcomes
- Expectations are not forecasts
- Country-level AI responses are not shown
What to watch next
- Publication of exact AI response tables
- AI-specific underwriting criteria
- Loan volumes and pricing
- Infrastructure borrower performance
- Power and customer-concentration tests
- Treatment of smaller software adopters
Living evidence record
Impact record IAI-1FVNAYU
Evidence stage
Observed
Confidence
Supported
Reporting basis
Source analysis
Independent or research support
Not yet
Record status
Monitoring
Last checked
27 October 2026
Source trail
1 direct source across 1 source type.
People impact
Documented in this record.
Uncertainty
Limits and next checks are explicit.
Stages describe the evidence available—not whether a technology is good or bad. See the public method.
Single-source reporting disclosure
This record analyses one direct source. It can establish what European Central Bank published or reported, but it is not independent corroboration of every performance claim or predicted outcome. The confidence label will change only when broader evidence is added.
Evidence trail
Sources used for this report
Links checked 27 October 2026
Research question, topics and regions
How euro-area banks describe current underwriting criteria and expected loan demand connected with AI adoption and infrastructure
Topics: Bank lending · AI infrastructure · Data centres · Credit risk · Corporate finance
Regions: Euro area · Europe
This report is labelled source analysis. We summarise and analyse source material in our own words; company statements remain attributed claims until independently supported. Translated summaries preserve the meaning of the original source and link back to it. Read our editorial standards.
Reader commentary
What does this mean where you are?
Tell us what you have seen at work, in your community or in everyday life. A question is welcome too. No account is needed; we review notes before publishing them.
Explore commentary across the portal →Published reader notes
0No published reader notes yet. You can start the evidence-led discussion above.
Prefer a private correction or response? Contact the newsroom.