Teradata collaborates with NVIDIA to enhance AI platform

Teradata has announced an exciting new collaboration with NVIDIA to enhance the Teradata Vantage platform using AI.

Teradata will integrate NVIDIA NeMo and NVIDIA NIM microservices into the Vantage platform, aiming to accelerate AI workloads and assist in developing foundational and customised large language models (LLMs), agentic workflows, and retrieval-augmented generation (RAG) applications. Customers will have the capability to deploy their own custom models via NVIDIA AI Enterprise, which provides enterprise-grade security, support, and stability, thereby facilitating a return on investment from generative AI use cases.

Beyond the software integrations, Teradata's platform will feature NVIDIA's accelerated computing infrastructure.

The initial implementation focuses on supporting both small language models (SLMs) and open LLMs within VantageCloud Lake. Through the new bring-your-own LLM capability, users can utilise NVIDIA AI accelerated computing platform clusters for LLM inferencing and model fine-tuning.

"Teradata customers are not casual users of their data and analytics, so we're thrilled to be engaged with NVIDIA to leverage our strong and trusted foundation for innovative AI use cases. By integrating NVIDIA's accelerated computing architecture into our Vantage platform, Teradata customers will be able to accelerate their use of AI at scale, in the environment of their choice, delivering unprecedented business value," stated Hillary Ashton, Chief Product Officer at Teradata.

NVIDIA NeMo Retriever, a collection of NVIDIA NIM microservices that allows organisations to connect custom models to diverse business data and deliver highly accurate responses, will also be available on Teradata VantageCloud. This will facilitate accelerated search and RAG applications to leverage existing customer data.

Teradata plans to provide access to NVIDIA NIM microservices, which are part of NVIDIA AI Enterprise, to VantageCloud customers before expanding this offering to hybrid customers globally.

"Data is the foundation of generative AI applications, enabling the development of today's highly customised applications. By integrating NVIDIA AI Enterprise, Teradata is providing developers a high-performance, full-stack platform that offers the security, stability and support enterprises require," Pat Lee, Vice President of Strategic Enterprise Partnerships at NVIDIA, said.

The Teradata VantageCloud Lake with NVIDIA AI accelerated computing will initially be available on AWS from November. Inference capabilities are set to be introduced in the fourth quarter, with fine-tuning availability expected in the first half of 2025. NVIDIA AI Enterprise integrations are anticipated to debut in 2025.