Snowflake Supercharges Machine Learning for Enterprises with Native Integration of NVIDIA CUDA-X Libraries

Snowflake announced a new integration with NVIDIA to accelerate ML workflows directly within Snowflake’s platform. Through the integration, Snowflake ML will now come preinstalled with some of NVIDIA’s most popular libraries for data science, offering Snowflake customers the ability to leverage GPU-accelerated algorithms for their ML workflows. This native integration simplifies and streamlines the entire ML model development lifecycle, allowing data scientists to accelerate model development for essential Python libraries, with no code changes required.

"Our vision is to help every company leverage data and AI with ease, security and performance, and this collaboration with NVIDIA helps us advance that goal," said Christian Kleinerman, EVP of Product, Snowflake. "By natively integrating NVIDIA CUDA-X libraries, we're giving our customers a massive performance boost. And this isn't just about faster performance; it's about enabling our data scientists to spend less time on infrastructure and more time deriving insights and achieving strong business outcomes for their organizations."

As enterprise datasets grow to unprecedented sizes, the need for GPU acceleration has become critical to maintaining productivity and managing costs. NVIDIA’s benchmark runs show speed up of 5x for Random Forest and up to 200x for HDBSCAN on NVIDIA A10 GPUs compared to CPUs. Through this integration, NVIDIA cuML and NVIDIA cuDF libraries – part of the NVIDIA CUDA-X Data Science (CUDA-X DS) ecosystem – are available in Snowflake ML to accelerate development cycles for scikit-learn, pandas, UMAP and HDBSCAN, without the need for code changes. 

"Data is the raw material of intelligence, and transforming it into insight is the foundation of generative and agentic AI,” said Pat Lee, VP of Strategic Enterprise Partnerships, NVIDIA. “By integrating NVIDIA cuDF and cuML libraries directly into the Snowflake ML platform, customers can now harness accelerated computing with their existing Python workflows, eliminating complexity and dramatically speeding up AI development."

The integration makes NVIDIA’s powerful CUDA-X Data Science (CUDA-X DS) ecosystem, an open-source suite of GPU-accelerated libraries, accessible directly through the Snowflake Container Runtime, a pre-built environment for large-scale machine learning development. Organizations now have the power to tackle computationally demanding challenges, such as:

The integration builds on Snowflake and NVIDIA’s continued collaboration to power generative AI capabilities within the AI Data Cloud. This new step further solidifies Snowflake’s commitment to deliver cutting-edge performance for all phases of the data and AI lifecycle. The companies will continue to work closely to provide Snowflake customers with seamless access to some of the most advanced GPU-accelerated tools, from traditional ML model development to the deployment of enterprise-grade LLMs.

Source: Snowflake media announcement