Teradata has announced the Autonomous Knowledge Platform, a new flagship product that unifies production-grade AI, analytics, and data into a single integrated system across cloud, on-premises, and hybrid environments. The platform is designed to operate autonomously, around the clock, without waiting to be asked — a shift from infrastructure built to respond to infrastructure built to act.
What it does
The platform turns structured and unstructured data, operating models, and experience into trusted, governed understanding. It provides business context for agentic AI to sense, decide, and act reliably and repeatedly across systems and tools with minimal human intervention, while learning and improving over time.
Key components include:
- Teradata AI Studio — A unified environment where users build, activate, and govern AI outcomes across the full lifecycle using analytics, ML, and agents. No switching tools, no exporting data, no rebuilding pipelines.
- Tera — Teradata's autonomous AI-powered workspace serving as a natural language interface with enterprise-grade agent execution environments. It includes built-in modes for data analysis (Tera Analyze), coding (Tera Code), and multi-agent system automation (Tera Claw).
- Tera Agents — Pre-built platform agents that perform tasks from continuously managing infrastructure to driving operational efficiency and cost optimization. New agents include: Sizing Agent (right-sizes compute resources dynamically), Telemetry Agent (observes platform signals), FinOps Agent (analyzes spend and consumption patterns), Tuning Agent (optimizes query execution), and Compute Agent (manages provisioning and concurrency).
- Teradata Cloud — The first available deployment, combining always-on Active Compute with on-demand Elastic Compute in a single managed system. It features a Connected Data Foundation with open table format support for Apache Iceberg and Delta Lake.
- Teradata Factory — Extends the platform on-premises with Dell PowerEdge servers, NVIDIA AI Infrastructure, NVIDIA AI Enterprise software, and high-performance networking for enterprises with strict data residency and regulatory requirements.
Key use cases
Teradata lists several primary use cases:
- Agentic Analytics — Business users access instant insights through natural-language queries, no SQL required.
- Hybrid Retrieval Agents — Production-ready agents built on the enterprise vector store, enabling billion-scale governed search across structured and unstructured data with time travel capabilities.
- End-to-End AI/ML Pipelines — Data scientists build, deploy, and operationalize scalable workflows with in-database analytics, no data movement required.
- Model Lifecycle Management — Includes ModelOps, Model Catalog, and LLM Ops spanning training, fine-tuning, monitoring, drift detection, and inference.
Partner integrations
Teradata is developing an ecosystem of partner integrations:
- Karini AI — Full-lifecycle, no-code agent development inside AI Studio.
- Pinecone — Low-latency vector retrieval for production workloads.
- Unstructured — Ingests unstructured data into the enterprise vector store, parsing 64+ file formats and 70+ connectors.
- WisdomAI — Agentic BI with natural-language queries, AI-powered dashboards, and proactive KPI monitoring.
Availability and pricing
The Teradata Autonomous Knowledge Platform is expected to be available in Q3 on Teradata Cloud. Teradata Factory availability will follow later this year. Tera Claw mode will be available in research preview by the end of the year. Teradata AI Services and AI Studio are available for all deployments now. Pricing details were not announced.
Bottom line
Teradata is positioning the Autonomous Knowledge Platform as the infrastructure layer for enterprises moving beyond AI pilots into production with autonomous agents. The platform addresses the tension between cost control and performance, cloud and on-premises, and experimentation and production stability — all within a single governed system. Whether it delivers on that promise will depend on real-world deployment at scale.