In a significant development at Snowflake Summit 2024, Snowflake has announced a new collaboration with NVIDIA to enable customers and partners to build customised AI applications on its platform. This collaboration brings together Snowflake's AI Data Cloud capabilities with NVIDIA's sophisticated AI-powered models and hardware, providing enterprises with streamlined options to deploy AI across their businesses.
The new partnership will see Snowflake integrate NVIDIA AI Enterprise software into its services, featuring the incorporation of NeMo Retriever microservices into Snowflake's Cortex AI. Cortex AI is Snowflake's fully managed large language model (LLM) and vector search service, which now benefits from enhanced connection capabilities to custom models, ensuring highly accurate and business-specific responses. Moreover, Snowflake Arctic, an enterprise-grade LLM, receives full support from NVIDIA TensorRT-LLM, further optimising its performance. Additionally, Arctic is now accessible as an NVIDIA NIM inference microservice, expanding its availability to more developers.
Sridhar Ramaswamy, CEO of Snowflake, remarked on the nature of the collaboration, saying, "Pairing NVIDIA's full stack accelerated computing and software with Snowflake's state-of-the-art AI capabilities in Cortex AI is game-changing. Together, we are unlocking a new era of AI where customers from every industry and every skill level can build custom AI applications on their enterprise data with ease, efficiency, and trust."
Jensen Huang, founder and CEO of NVIDIA, added, "Data is the essential raw material of the AI industrial revolution. Together, NVIDIA and Snowflake will help enterprises refine their proprietary business data and transform it into valuable generative AI."
NVIDIA AI Enterprise software's key capabilities, such as NeMo Retriever, are integrated into Snowflake's Cortex AI, increasing the accuracy and performance of AI applications aimed at retrieval-augmented generation-based solutions. Another notable addition is the NVIDIA Triton Inference Server, which supports deploying, running, and scaling AI inference for various applications across different platforms.
Furthermore, NVIDIA NIM inference microservices, part of the NVIDIA AI Enterprise suite, can now be deployed within Snowflake via Snowpark Container Services. This feature allows organisations to deploy a series of foundational models directly within the Snowflake ecosystem, contributing to secure data processing and faster time-to-value.
Quantiphi, an AI-centric digital engineering company and elite partner of both Snowflake and NVIDIA, is among the many firms leveraging this new capability. Quantiphi's native applications, such as "baioniq" - a generative AI platform to enhance knowledge worker productivity, and "Dociphi", an AI-driven document processing platform for the banking, financial services, and insurance sectors, are developed using the NVIDIA NeMo framework and will be available on Snowflake Marketplace for seamless user deployment within the Snowflake environment.
Additionally, Snowflake Arctic LLM, launched in April 2024 and trained on NVIDIA H100 Tensor Core GPUs, is now offered as an NVIDIA NIM. This provides users with instant access to the LLM and the flexibility to host the Arctic NIM on their preferred infrastructure. This expansion allows developers even more choices in deploying the comprehensive enterprise LLM.
The recent announcements build upon Snowflake and NVIDIA's earlier collaborations to provide a unified AI infrastructure and computing platform within the AI Data Cloud. These advancements mark pivotal progress in Snowflake and NVIDIA's shared mission to support enterprises on their AI journeys.
