Data platforms are the analytical core of an organisation: lakehouses, cloud data warehouses, data-lake storage and analytical databases where data from transactional systems, sensors and SaaS tools comes together. Teams use them for reporting, for training models and for sharing datasets with partners. Well-known non-European reference points are Snowflake and Databricks.
The risk lies in the combination of volume and entanglement. Storage, compute and metadata are often tied together, so a move also means rewriting queries, pipelines and access rules. Customer-held keys only mean something if the vendor cannot bypass them and logs, caches and backups are covered too. The Data Act requires providers of data processing services to make switching easier, with open interfaces, export in commonly used machine-readable formats and removal of switching charges, including data egress fees. It also sets safeguards against unlawful third-country government access to non-personal data.
When evaluating a European option, look at open table formats such as Parquet or Iceberg, standard SQL, separation of storage from compute, and whether you can also run the platform yourself. Have the practical exit procedure and its cost written into the contract.