How Raiffeisen Bank International Cut Analytics TCO by 5x with Databricks SQL

Raiffeisen Bank International operates a highly federated network of banking subsidiaries across Central and Eastern Europe, serving retail, corporate, and institutional customers under stringent regulatory requirements for security, governance, and auditability. Over decades of organic growth and acquisition, the group accumulated a fragmented analytics environment: each bank and department had built its own SQL conventions, access controls, and operational models. Collaboration across teams was difficult, code reuse was limited, and there was no consistent way to govern or audit data usage across the group.

The result was a patchwork of on-premises and cloud systems, each with different dialects and operational models. As George Moldovan, the data leader who oversaw the transformation, described it: legacy SQL conventions and systems developed over time had become insufficiently strong, unified, or modern enough to support the data- and AI-driven demands of contemporary banking. The gap between what the analytics infrastructure could deliver and what risk, compliance, finance, and retail teams needed was widening.

RBI chose Databricks SQL as the shared analytics foundation for a new internal platform called APEX. Rather than a risky big-bang cutover, the team ran legacy platforms and Databricks in parallel, onboarding users gradually, validating results, and protecting production workloads throughout. Moldovan was clear that migration is fundamentally a change management problem — onboarding, governance, and building trust in the new platform — not a technical one. The team invested accordingly.

Once workloads began running on Databricks SQL, the performance improvements were immediate. Queries that had constrained analysts due to long runtimes became interactive. In one extreme case, a query that took 30 days on the legacy infrastructure completed in 12 minutes on Databricks. Average workloads now run 3–4x faster than on the previous cloud solution. Elastic compute allows teams to scale resources on demand for large analytical jobs, supporting hundreds of concurrent users across risk, compliance, and finance without performance degradation.

Cost improvement came from architectural simplification rather than usage restrictions. By consolidating analytics onto Databricks SQL and serving BI workloads directly from the platform, RBI significantly reduced unnecessary data movement — historically one of the largest drivers of cloud analytics cost. The bank built its own cost-monitoring and forecasting platform on Databricks to track usage by warehouse, user, and workload. The resulting TCO is 5x lower than the previous cloud solution. For a European banking group, governance and open standards were equally important: Databricks' support for open formats and APIs ensures that RBI's data foundation remains portable and regulatorily transparent as requirements continue to evolve.