Brief
The essential idea
Revolut's 2025-2027 strategy couples aggressive geographic expansion with a target of 100 million customers by mid-2027, new licenses in Latin America, India, and Western Europe, and deeper use of AI. The operating model uses small squads of roughly six to eight people, end-to-end product ownership, and developer responsibility for quality rather than a separate QA handoff.
The technology stack is cloud-native on GCP, using GKE and Compute Engine, Java 21, PostgreSQL with jOOQ, Redis, Kubernetes, domain-driven design, CQRS, test-driven development, and event-driven integration. Regulatory and compliance needs are treated as architectural constraints, not review steps added after delivery.
The distinctive leverage comes from Tower and Rhea. Tower provides a service catalog, ownership, SLOs, and approved patterns; Rhea uses a GitOps model to generate pipelines, secrets, cloud resources, alerts, and dashboards. This extreme automation increases speed and consistency, but also centralizes risk and can narrow team freedom, so the platform must stay responsive to product-team needs.
Decision lens
Key takeaways
Small squads move quickly when they own product outcomes, delivery, and production quality end to end.
A service catalog makes ownership, SLOs, and operational expectations discoverable.
GitOps automation can provision delivery, infrastructure, secrets, and observability as one repeatable path.
Compliance is most scalable when encoded into platform defaults and architecture.
Extreme standardization trades local freedom for organizational speed and risk control.
AI adoption is stronger when embedded in the operating model instead of isolated in a lab.
Workplace experiment
Apply it at work
- 1
Audit whether every production service has a discoverable owner, SLO, runbook, and approved delivery path.
- 2
Automate one repeated provisioning workflow across code, infrastructure, secrets, alerts, and dashboards.
- 3
Define which platform standards are mandatory guardrails and where squads retain explicit freedom.
- 4
Review whether quality ownership sits inside the squad or is delayed by a separate handoff.
Choose one action, define the observable effect, and keep the first test small enough to reverse.
Evidence
Sources and further reading
Additional sources
Channel, aggregator, and commentary links confirm the work; they are not the primary source.
Local knowledge map
A small, typed neighborhood instead of the full-catalog graph.