Barclays and DNB Cut IT Jobs; Mizuho Takes the Opposite Route

NatWest eliminated approximately 1,000 full-time positions in the third quarter, according to results released last week, totaling 1,400 cuts for the year. On Friday (30), CaixaBank's CEO, Gonzalo Gortázar, stated during the results presentation up to September that artificial intelligence will not lead to job cuts at the bank in the short term. These two announcements, made days apart, are usually interpreted as opposing sides of a ledger. The comparison, however, is misguided: the useful question is not how many jobs AI eliminates, but which ones.

The Primary Target is Technology Itself

The job cuts attributed to AI in Europe are disproportionately affecting support and technology areas rather than branch operations. On October 6, Norwegian bank DNB reported to the Oslo Stock Exchange that it would cut around 400 positions in its Technology & Services unit by the end of 2026, nearly 4% of a workforce of over 10,500. The bank cited AI agents' effectiveness in managing customer data, KYC processes, technology development, and programming.

On October 22, Barclays announced around 5,000 layoffs by the end of the year, which is about 5% of its workforce. According to CEO C. S. Venkatakrishnan, 60% of the cuts will come from support roles, with fewer developers needed as AI makes each professional more productive. At NatWest, cost savings reached around £420 million.

The Case Against the Thesis

Gortázar is the clearest voice on the other side. He argues that efficiency should not be confused with fewer employees, and the bank can achieve more with the same team. CaixaBank, which reported a profit of €4.962 billion over nine months, a 12.9% increase, expanded its CaixaBank Tech subsidiary from 300 to 1,700 people. He acknowledges that some sectors may end up with excess staff but does not foresee this for the bank in the short term, a caveat that should not go unnoticed.

There is evidence that undermines the reading that AI is already making mass cuts. NatWest's overall headcount actually rose in the quarter because the acquisition of Evelyn Partners added around 2,200 employees. Venkatakrishnan himself downplays the role of technology: he referred to AI as "the icing on the cake, not the cake," attributing gains to years of investment in technology. Treating each layoff as proof that AI replaces bank workers is as misleading as treating CaixaBank's rhetoric as evidence that it does not.

Japan: Retraining Instead of Cutting

Mizuho Financial Group will launch the Mizuho AI Bootcamp in November, according to Nikkei and other Japanese outlets that reported on the plan. The program aims to transform sales and back-office employees into AI application developers, utilizing assisted programming tools to reduce training time from around a year and a half to three months. In the pilot program, between June and August, five participants—four without prior experience—passed the assessment and moved into engineering roles in September. The initial goal is about 100 individuals, selected from the approximately 50,000 employees in the group, with the aim of reaching several thousand in ten years.

The difference stems from the labor market. In Japan, where the active population is declining and labor laws make layoffs costly, the bottleneck is a lack of technical personnel, not an excess of staff. In the UK and Norway, banks are using the same tool, AI-assisted programming, to reduce the number of developers. In Spain, CaixaBank uses the gains to expand its technology division. The technology is the same in all three cases; it is the decision on how to utilize the gains that varies.

What This Means for Service Providers to Banks

The distinction that the shallow debate overlooks is between cutting technology jobs and cutting the bank itself. DNB and Barclays are reducing exactly the type of roles that, a decade ago, were outsourced to consultancies and integrators: programming, data maintenance, operational KYC. If these tasks shrink within the bank due to AI agents, the next question is whether they will also diminish in contracts with external providers performing the same work at delivery centers outside Europe.

The Mizuho case points to an alternative answer. A bank transforming 100 back-office employees into developers in three months is internalizing capabilities it would previously have outsourced. In either model, the market for those selling development hours to banks diminishes. What changes between London, Oslo, Barcelona, and Tokyo is merely whether the gains go to shareholders or to the retrained employee.