Teradata launches AI Factory for secure, cost-predictable AI
Teradata has unveiled the Teradata AI Factory, an on-premises solution designed to support enterprises in developing and operationalising artificial intelligence with a focus on security, compliance, and predictable costs.
The Teradata AI Factory brings together core Teradata technologies and selected third-party components, such as those from NVIDIA, within a single system configured for AI development and workflows. The product's integration aims to address the increasing complexity of AI ecosystems and enterprises' needs for regulatory compliance, especially around data sovereignty and local data storage requirements set by global regulations such as GDPR and HIPAA.
Addressing governance and cost control
Teradata AI Factory merges data pipelines, algorithm execution environments, and software infrastructure into an integrated stack. According to Teradata, the ready-to-run system is suitable for sectors with heightened regulatory demands, including healthcare, finance, and government, and is also aimed at enterprises wishing for greater control over their AI deployments.
The AI Factory incorporates Teradata AI Microservices and compatibility with NVIDIA's hardware, including support for NVIDIA GPUs supplied by customers. This collaboration is intended to facilitate faster AI model development, with specific emphasis on native retrieval-augmented generation (RAG) processing to enhance performance and reliability for enterprise-scale projects.
Market trends indicate a shift towards greater on-premises AI deployment. Referencing a recent Gartner report, Teradata noted, "By 2028, more than 20% of enterprises will run AI workloads (training or inference) locally in their data centres, an increase from approximately 2% as of early 2025." (Gartner, "How to Determine Infrastructure Requirements for On-Premises Generation AI," March 2025).
Security, compliance, and control
As regulatory scrutiny rises and the cost of cloud services remains unpredictable, businesses are increasingly motivated to adopt private AI solutions that keep data within organisational boundaries. Locally processing sensitive data mitigates risks related to data exposure, intellectual property, and non-compliance.
Teradata AI Factory addresses these concerns by offering an infrastructure where data remains on-premises, aligning with compliance demands and reducing vulnerabilities often associated with public or shared platforms. This setup allows for adherence to standards such as GDPR and HIPAA, strengthening data privacy and governance.
"Market dynamics are increasing buyer interest in on-premises solutions. Teradata remains the clear leader in this environment, with proven foundations in what makes AI meaningful and trustworthy: Top-notch speed (performance), predictable cost (resource efficiency), and integration with the golden data record (which may already live on Teradata). Teradata AI Factory builds on these strengths in a single solution for organisations using on-prem infrastructure to gain control, meet sovereignty needs, and accelerate AI ROI," said Teradata's Chief Product Officer, Sumeet Arora.
The solution is also designed to help enterprises optimise existing investments in GPUs and on-premises hardware, avoiding the unpredictability of cloud-based resource pricing. This is achieved through a localised architecture that reduces data movement and latency, aiming for consistent operational expenses and longer-term financial efficiency.
Integrated tools and platform features
The AI Factory includes Teradata's Enterprise Vector Store to handle both structured and unstructured data for generative AI use cases. A new AI Workbench provides self-service access to analytics libraries and tools, supporting model lifecycle management, compliance checking, and deployment of large language models. Supported developer tools include JupyterHub, ModelOps, Airflow, Gitea, and Devpi—all intended to streamline deployment and reduce technical friction.
The platform is built on Teradata's IntelliFlex system, which integrates with customer-provided GPUs and supports high-performance AI workloads. AI Microservices with NVIDIA NeMo underpin native RAG pipeline capabilities for natural language querying and retrieval tasks at scale. The solution also includes modern data ingestion tools, compatibility with open table formats and object stores, and integration with NVIDIA utilities for handling various data types, including PDFs.
With the launch of AI Factory, Teradata is positioning itself as a resource for enterprises needing comprehensive, in-house AI capability with security and governance features suited to modern regulatory and operational requirements.
