The implemented project and the attribution boundary
This independent analysis examines BBVA's publicly documented ADA migration to AWS. XIVTech did not deliver the project, and BBVA is not presented as an XIVTech client. The discussion below distinguishes facts reported by the project organizations from our technical interpretation.
BBVA's December 26, 2024 announcement states that migration to its Analytics, Data, AI platform had been completed for Europe and Uruguay. Data Center Dynamics reported that milestone on January 9, 2025. The detailed AWS and BBVA engineering retrospective published in October 2025 covers a broader Europe-and-Americas migration scope. These are different dated observations, not one measurement window.
BBVA's December 2024 Europe and Uruguay completion announcement
An inventory of things that execute, not only things stored
The October 2025 methodology describes moving approximately 500 TB of data and more than 250 analytical environments, alongside job configurations, executable artifacts from Artifactory, internal libraries, notebooks, BI data and analytical developments. It also describes adapting input/output paths and running parallel validations.
Our interpretation is that the executable dependency boundary is more useful than a storage-only boundary when selecting migration work. Copying data cannot establish that the job configuration, library behavior and permissions required to use it have arrived correctly. This is a lesson drawn from the published case, not a claim about an undocumented BBVA checklist.
AWS and BBVA's October 2025 migration methodology
Copied objects and a usable catalog are different results
The methodology describes a modified DistCp process on Amazon EMR for HDFS-to-S3 copying and a separate catalog and partition synchronization process. The team reused the copy inventory to identify partition changes and interacted with AWS Glue APIs through orchestration and Lambda functions.
