Teradata

Teradata is an enterprise data warehouse and AI platform that runs in Teradata Cloud, a customer's own cloud or on premises, sold on consumption pricing.

Teradata gives large organisations one place to keep governed data, analyse it at scale and let AI agents act on the results. Teradata Corporation, a publicly traded company headquartered in San Diego, California, builds and sells it. The company presents the platform as the Teradata Autonomous Knowledge Platform, designed for a time when agents query enterprise data alongside analysts. One platform runs in Teradata Cloud, inside a customer's own cloud account or on premises, including sovereign and air-gapped setups. The Foundation layer stores data and runs workloads, with compute clusters that scale apart from storage and grow automatically under agent or peak analytical demand. Open table format support works with Apache Iceberg and Delta Lake tables held in AWS S3, Azure Data Lake Storage or Google Cloud Storage. Those tables offer time travel and rollback to earlier versions. Row-level security constraints filter what each user role can see. Python runs inside the database through the SCRIPT table operator and the teradataml package, next to ClearScape Analytics functions. Above the foundation, Knowledge Context turns multi-modal data into context for AI, AI Studio builds and manages AI work, and Agent Execution secures and controls multi-agent systems. Teradata suits enterprises whose analytics are mission-critical, in retail, financial services, healthcare, life sciences, manufacturing, telecommunications, transportation and public services. It fits a data team with a steady analytics baseline that also needs room for AI-agent load and peak demand. Regulated organisations that must keep data and models under their own control can run Factory, the private on-premises option. Teams moving to open table formats can keep one copy of their data for several engines. Analysts point Tableau at warehouse tables for reporting, while data engineers schedule loads and transformations with dbt, Airflow or Dagster. Lakehouse teams leave Iceberg and Delta tables in their cloud object store and query them where they sit. Data scientists train and score many models without moving rows out of the database. AI teams ground their agents in the same governed tables the business already reports from, so an agent's actions rest on trusted numbers.