S2 E5 | Beyond Hyperscalers: How to Run Modern Data Platforms on European Clouds

EU clouds without the hype. Niels (Partner & Lead Data Engineer at Dataminded, and our technical hiring lead) breaks down data sovereignty vs. Cloud Act, GDPR realities, and a portable, Kubernetes-first stack with Iceberg, Trino, and Airflow. We compare Scaleway, OVH, Exoscale, UpCloud, look at cost drivers, encryption/KMS, egress policies, and how to avoid vendor lock-in plus when best-of-breed beats all-in-one and why “keep it simple” still wins.

What you’ll learn

• When EU clouds make more sense than hyperscalers (and when they don’t)

• Designing a portable platform: Terraform/Tofu for infra, Argo CD for apps

• Table formats 101: why Apache Iceberg over plain Parquet/CSV

• Query layer choices: Trino for open SQL across object storage & DBs

• Orchestration in practice: Airflow patterns, dependencies, SLAs

• Security & governance: OPA for fine-grained policies, IAM, catalogs

• Cost & ops: egress, managed services gaps, version lag, troubleshooting

• Team skills: what to hire for, and the “hard questions” Niels asks in interviews

Topics: EU cloud, European cloud providers, data sovereignty, GDPR, Cloud Act, Kubernetes data platform, Apache Iceberg, Trino, Airflow, vendor lock-in, OPA, Argo CD, Terraform, Exoscale, Scaleway, OVH, UpCloud