How Validio is building the compliance layer for enterprise AI

For high-risk AI systems, the gap between current data practices and what regulators require is becoming harder to ignore: Article 10 of the EU AI Act, enforceable from August 2026, requires that training, validation, and testing datasets be representative and free of errors, with documented governance practices covering from data collection to bias detection.

Validio (Sweden, founded 2019), co-founded by Patrik Liu Tran, Oliver Molander, and Urban Eriksson, was built to address this at the infrastructure level. The platform monitors enterprise data pipelines in real time, using automated anomaly detection across large datasets to catch quality failures before they reach production AI systems. When a failure is detected, it traces the origin through integrated data lineage and cataloguing, generating an auditable record of how data has moved and been transformed across systems. The platform also integrates with cloud data platforms, giving engineering and compliance teams a unified view of data health across the full pipeline - including automated quality checks and documentation outputs relevant to regulatory reporting.

In March 2026, Validio raised $30 million in a Series A round led by Plural, with continued participation from Lakestar and J12 Ventures and other angel investors. Total funding stands at $47 million, and the capital will support expansion into the US, UK, and Northern Europe.

Sources: Validio | EU AI Act Article 10 | EU AI Act Article 12 Founders: Patrik Liu Tran, Oliver Molander, Urban Eriksson