In a strategic move to address the increasing demands for data sovereignty, predictable costs, and total control over artificial intelligence processes, Teradata has announced the launch of Teradata AI Factory, an integrated “ready-to-use” solution that brings the power of generative, predictive, and agentic AI to on-premises environments, especially in highly regulated sectors.
Developed in collaboration with NVIDIA, this platform offers a robust combination of software, hardware, and services designed to accelerate business innovation through private and reliable artificial intelligence.
What is Teradata AI Factory?
This is a solution designed to run within the client’s own facilities, allowing for complete control of data while avoiding the risks of exposure in public or shared environments. AI Factory integrates essential components such as data pipelines, algorithm execution, software infrastructure, and third-party AI tools, all within a single scalable system.
Thanks to its integration with Teradata AI Microservices and customer-provided NVIDIA GPUs, AI Factory enables the efficient and secure execution of advanced models – including RAG (Retrieval-Augmented Generation). This architecture, based on NVIDIA Enterprise AI Factory’s design validation, is designed for industries such as healthcare, finance, and the public sector.
A Response to Market Tensions in AI
In a context marked by global instability, regulatory pressure, and rising cloud service costs, more organizations are seeking local solutions for running AI workloads. According to Gartner, by 2028, more than 20% of companies will run AI workloads in their own data centers, up from a mere 2% today.
“Customers want to control their costs, protect their sensitive data, and accelerate their ROI in AI,” stated Sumeet Arora, Chief Product Officer of Teradata. “With AI Factory, we offer a comprehensive solution built on our leadership in performance, efficiency, and data governance.”
Key Benefits of Teradata AI Factory
- Maximum Security and Regulatory Compliance
Sensitive data does not leave the organization, minimizing exposure risks and ensuring compliance with regulations like GDPR and HIPAA. It is ideal for environments where data sovereignty is critical. - Efficient Performance and Predictable Costs
The on-premises architecture reduces latency and avoids unpredictable cloud costs. AI can be executed on both CPUs and with acceleration using proprietary GPUs, maximizing hardware utilization. - Ready to Use, No Complexities
AI Factory comes with an integrated stack that includes modern tools like JupyterHub, Airflow, Gitea, or Devpi, as well as modeling features, one-click LLM deployment, and model lifecycle management. - Integrated Generative AI
Thanks to NVIDIA NeMo and Teradata’s microservices, it includes native capabilities for RAG, allowing natural language queries over structured and unstructured data, generating intelligent and scalable insights.
System Components
- AI IntelliFlex Platform: hardware base optimized for AI workloads.
- AI Workbench: work environment with access to libraries and analysis tools.
- Enterprise Vector Store: for merging structured and unstructured data in generative AI.
- Modern Data Pipelines: support for open formats, object stores, and complex documents like PDFs.
Now Available
Teradata AI Factory is now available for companies seeking technological autonomy and scalable, secure, and efficient AI deployments. With this solution, Teradata strengthens its position as a key provider in the convergence of advanced analytics, data management, and next-generation AI capabilities.