Rabobank said on August 4 that it will invest up to 2 billion euros over the next three years in strengthening its data and IT foundation, improving digital customer experience, and scaling artificial intelligence, an announcement the Dutch cooperative bank paired with first-half 2026 results showing net profit held essentially flat at 2.694 billion euros.
The bank’s CET1 capital ratio stood at 20.2%, and its cost-to-income ratio improved to 50.2%, giving it room to fund the AI push from a position of strength rather than as a defensive cost cut. CEO Stefaan Decraene tied the spending directly to customer expectations and competitive pressure, saying the bank needs the investment “to take full advantage of these developments” in AI and to meet “evolving customer expectations.” The commitment builds on Rabobank’s Agentic Hub, launched in June, which centralizes best practices and pre-built agentic AI tools for rollout across the bank’s roughly 1,100 software development teams.
What stands out is the architecture choice behind the number. Rather than centralizing AI development in one team, the way HSBC concentrated its AI work into a single global center of excellence, Rabobank is funding a distributed model that pushes agentic tools out to every one of its development teams, coordinated through the Agentic Hub rather than a single control point. That is a harder governance problem: a distributed rollout across 1,100 teams means more places for model risk, audit trails, and access controls to diverge, which is exactly the kind of operational risk question a bank’s compliance function will need to get ahead of before, not after, the tools are already embedded in 1,100 codebases.
Source: Rabobank
