Singapore – Teradata has introduced new privacy, security, and resiliency capabilities designed to protect the sensitive data that enterprise AI systems depend on, according to the company.
The San Diego-based firm announced the update on 20 October 2026, describing it as an integrated approach that reduces raw identity exposure, enforces protection closer to the data and keeps AI-related data recoverable.
Teradata argues that AI changes the risk profile of enterprise data. The company says AI is built to combine disparate information, so data that is not individually identifiable can become identifiable once a model brings it together.
At the same time, Teradata notes that sensitive information now moves among users, applications, models and agents, which makes securing the boundary alone more difficult.
Sumeet Arora, Chief Product Officer, Teradata, said, “Privacy, security, and resiliency have historically been treated as separate problems with separate solutions. That approach doesn’t hold up when AI is moving sensitive data across users, applications, models, and agents simultaneously.”
“This announcement reflects how we think those capabilities need to work together, and it’s the foundation for how we intend to extend that protection as AI environments grow more complex and the actors operating within them multiply,” Arora added.
On privacy, Teradata has added Poly-Anonymization®, which it delivers through an OEM partnership with data security firm Anonomatic. The feature replaces personally identifiable information (PII) fields with a stable identifier, while the underlying identity is held separately in Anonomatic PII Vault, which Teradata describes as a post-quantum-secure environment.
Authorised users can restore identity when a legitimate use case requires it, whereas other users work with data that keeps its analytical structure and relationships. Teradata says this allows sensitive data to support AI, analytics, testing and approved collaboration without broadly exposing who the data is about.
The company contrasts the approach with methods that rely solely on mathematically reversible encryption. It says PII Vault combines identity separation, Poly-Anonymization, access controls and additional layered protections instead.
Teradata has also added format-preserving encryption for cases where a value’s structure matters to downstream systems. The technique protects sensitive values without changing their format, so connected applications and workflows continue to function.
Teradata reports that large-scale testing of the Anonomatic-powered controls showed no performance penalty for enterprise analytical workloads. In those tests, multi-table join queries on protected data used fewer Teradata resources than equivalent queries on unprotected data, although the company says results will vary by workload and configuration.
Matt Fleck, Founder and CEO of Anonomatic, a data security software company focused on making PII data privacy secure, simple and cost effective, said, “Enterprise AI creates a fundamentally different privacy challenge. Organizations need to use sensitive, individual-level data to build effective AI, but the traditional tools for protecting that data weren’t designed to preserve what AI needs to work.”
“Our partnership with Teradata makes it possible to reduce identity exposure at enterprise scale without forcing organizations to choose between protection and utility,” Fleck continued.
Beyond privacy, Teradata is applying classification, policy enforcement and access controls closer to the data itself, rather than only at the surrounding systems. New native capabilities include dynamic data masking, data classification and tagging, column-level encryption and global identity.
For data in transit, the company has introduced post-quantum protection through TLS 1.3 hybrid-key exchange, which combines classical and post-quantum cryptographic algorithms in a single handshake. Teradata says advances in quantum computing are making current encryption standards increasingly vulnerable.
Through a partnership with Immuta, a data authorisation platform that sits between identity and enterprise data, Teradata is also offering dynamic attribute-based access control (ABAC). The company says this enforces policy consistently across users and AI workloads without duplicating data or rewriting governance rules.
On resiliency, Teradata Enterprise Data Protection lets organisations match protection levels to business criticality, instead of applying the same architecture to every workload. The company says the aim is to restore the trusted context AI needs to resume mission-critical operations, not simply the data itself.
Database-level high availability, delivered through real-time replication, adds an availability layer built for continuous synchronisation and low recovery objectives. Teradata distinguishes it from disaster recovery.
Immutable backup protection, meanwhile, ensures recovery points cannot be modified, encrypted or deleted, even by privileged administrators. Teradata points to ransomware increasingly targeting backup data as the reason for the feature.
Backup and recovery, scheduled data synchronisation and disaster recovery complete the Enterprise Data Protection set.
Looking ahead, Teradata says both its security and resiliency architectures are designed to evolve, including through continued investment in controls that govern how AI agents access and act on enterprise data within policy boundaries.
