Summary

• Euro area firms are expected to allocate around 10% of investment to AI in 2026, with AI-related lending accounting for approximately a quarter of recent corporate credit growth.
• ECB surveys find that 52% of workers use AI at work, although significant business adoption and measurable economy-wide productivity gains remain limited.
• Philip Lane warns that changing employment expectations may affect wage bargaining, while global AI investment could contribute to higher long-term borrowing costs.

Artificial intelligence is already influencing European business investment and borrowing costs, while expectations about automation may be affecting workers’ bargaining power before the technology has produced a measurable increase in economy-wide productivity, according to European Central Bank chief economist Philip Lane.

Speaking at the International AI Summit in Dublin on 13 October, Lane described a euro area economy in which AI adoption is advancing faster than the evidence of its economic benefits. Firms are spending more on software, computing infrastructure and employee training, while financial markets are responding to the global investment required to develop advanced AI systems.

The ECB estimates that euro area firms will devote around 10% of their total investment to AI in 2026, while AI-related lending accounts for approximately a quarter of recent growth in credit to businesses. Those figures indicate that the technology is influencing economic activity even though the bank has not yet identified a material effect on aggregate productivity.

Lane also raised the possibility that expectations of automation could weaken workers’ negotiating position. Employment levels have not shown a major AI-related change, but subdued hiring plans and uncertainty about future demand for some occupations may influence wage negotiations before significant displacement appears in labour market statistics.

Adoption exceeds measurable productivity gains

ECB surveys indicate that AI is becoming part of working life across the euro area, although intensity of use differs substantially between businesses. In the fourth quarter of 2025, 64% of firms responding to the Survey on the Access to Finance of Enterprises reported using AI experimentally or moderately, while only 7% described their use as significant.

The distinction indicates that access to AI tools is more widespread than their integration into core operating processes. Employees may use publicly available assistants for individual tasks without their employer having redesigned workflows or made substantial investment in specialised systems.

Among workers, the ECB’s household survey shows that the proportion reporting AI use at work increased from 26% in 2024 to 52% in 2026. Respondents also report time savings, particularly in activities such as software development, although those improvements are concentrated in particular occupations and tasks.

Time saved during an individual activity does not necessarily produce an equivalent improvement in an organisation’s output. AI-generated work may require checking, while other stages of a process can continue to limit the number of services or products a business delivers.

Meaningful productivity gains may consequently depend on investment in systems beyond the AI application itself. Businesses need suitable data, reliable integration with existing software and employees capable of using the technology effectively, while managers may have to reorganise responsibilities before gains become visible in financial results.

The ECB’s survey of planned investment reflects those requirements. Among firms intending to spend on AI, 49% identified technologies and tools, 46% employee training, 40% data and infrastructure, and 12% specialist recruitment. Respondents could select several categories, and the percentages describe intentions rather than completed investment.

The central bank has also found that businesses expect to fund much of this expenditure through their own resources. That may reduce immediate dependence on external borrowing for smaller adoption projects, although major infrastructure developments require different financing arrangements and can have a greater effect on financial markets.

Employment expectations enter wage negotiations

The implications for workers are more difficult to measure because businesses can adjust employment gradually. A company expecting software to automate some tasks might reduce recruitment, delay replacing departing employees or alter the responsibilities of existing staff without announcing substantial redundancies.

Such decisions can affect graduates particularly strongly because entry-level roles often provide the initial experience needed to enter professional occupations. Lane noted subdued hiring intentions and weaker prospects for new graduates, while stopping short of attributing the entire change to AI.

Workers already employed may also respond to expectations about future demand. If an occupation appears increasingly exposed to automation, employees could become less willing to negotiate aggressively over pay or move between employers, particularly where alternative opportunities are uncertain.

Lane identified this as a possible channel through which AI could moderate wage growth and reduce labour’s share of income. The argument concerns how employment expectations influence bargaining, rather than a finding that AI has already caused a measurable decline in wages across Europe.

Other economic conditions continue to influence recruitment and pay decisions, including demand for goods and services, energy prices and financing costs. The ECB’s employment evidence therefore cannot isolate the effect of AI from all other changes affecting the labour market.

Occupational differences also matter. Companies developing AI infrastructure and supplying associated equipment may require additional engineers and technical specialists, while some knowledge-intensive businesses see greater potential for software to substitute for existing tasks.

These differences are reflected in the ECB’s surveys, which find more optimistic expectations among workers in sectors benefiting from AI infrastructure investment than among those facing greater automation risk.

Techopia previously examined hiring trends among Danish businesses adopting AI, where changes in recruitment provided an earlier indication of adjustment than widespread job losses. Lane’s remarks place those developments within the broader relationship between employment expectations, wages and monetary conditions.

Global investment influences European interest rates

Beyond its effects on business processes, AI is becoming an important influence on capital markets. Training advanced models and operating large computing facilities requires substantial investment in processors, data centres, electrical infrastructure and specialist equipment, with financing increasingly extending into corporate debt markets.

Lane suggested that the global surge in AI investment, together with expectations of future productivity growth, may be contributing to higher long-term interest rates. These financing conditions can affect European companies even when they operate outside the AI sector.

Higher market borrowing costs can make investment more expensive, while changes in financial conditions can influence economic demand and the outlook for inflation. The ECB must assess those effects alongside the possibility that productivity improvements eventually increase the economy’s capacity to produce goods and services.

Lane did not present AI as the sole cause of rising long-term yields or announce a change in the ECB’s policy rate. His assessment identifies a possible influence on financing conditions that the central bank needs to consider as part of its wider monetary analysis.

The international distribution of AI investment complicates the European picture. Lane cited research indicating that the United States hosts approximately 75% of global AI computing capacity, compared with around 5% in Europe, while US digital investment has increased faster than equivalent spending in the euro area.

European companies can obtain AI services from international providers and participate in supply chains serving the global infrastructure expansion. However, the location of computing capacity and ownership of technology can affect where investment occurs and how much of the associated economic activity remains within Europe.

Large AI valuations and growing borrowing also create financial stability concerns. A reversal in expectations about technological progress or commercial returns could affect share prices and credit markets, particularly where investment depends heavily on future revenue growth.

For the ECB, AI is therefore influencing several economic variables on different timescales. Investment and borrowing activity are already measurable, while substantial productivity gains and labour market changes remain uncertain. The bank’s policy assessment will depend on how those developments appear in subsequent data, rather than assuming that widespread use of AI immediately translates into higher output.