Snowflake introduced new feature updates to its AI data cloud platform at the Summit 2024 event. Snowflake’s enterprise AI simplifies all development workflows and decision-making, from data analysis to AI/ML capabilities and application building, to support operational innovation in enterprises.

Snowflake announced new innovations for the “Era of Enterprise AI” at its “Snowflake Summit 2024” event.

The main announcements made by Snowflake at this event were the introduction of the ‘Polaris Catalog’ and strengthening of its partnership with NVIDIA.

Snowflake is working with NVIDIA to enable customers to build custom AI data applications based on NVIDIA AI on its platform. That is, it integrates NVIDIA's 'NeMo Retriever' microservice and Snowflake's 'Cortex AI'.

Snowflake CEO Sridhar Ramaswamy and NVIDIA CEO Jensen Huang emphasized that this collaboration will “make it easier and more efficient for customers to seamlessly connect custom models to diverse business data and build custom AI applications.”

Snowflake announced new features to strengthen its data foundation. “A strong data foundation is essential for building custom AI applications,” said Choi Ki-young, CEO of Snowflake. “You need a data foundation that is secure and has unified governance.”

‘Polaris Catalog’, recently released by Snowflake, is a new open source-based data catalog service. It was developed with the 'Apache Iceberg' specification to improve data interoperability and flexibility between heterogeneous systems.

This demonstrates Snowflake’s commitment to open source by not only enabling large enterprise customers to integrate their data in open formats, but also by providing access at the catalog layer to choose their compute engine.

Snowflake explained, "Users can utilize Iceberg Tables with interoperability across a variety of cloud service platforms, including AWS, Google Cloud, and Microsoft Azure, and can freely replace hosting infrastructure, so they are not locked into a specific provider."

Meanwhile, Snowflake announced new features for Cortex AI, its industry-leading enterprise AI. Cortex AI is a tool for easily and efficiently creating generative AI-based applications. It unveiled two new chat features: Snowflake Cortex Analyst, and Snowflake Cortex Search, an enterprise-grade hybrid search function.

“Cortex Analyst is a tool for building generative AI-based applications,” said Snowflake CEO Jo Seong-hyun. "Although implementing 'Text-to-SQL', where LLM creates SQL statements, is more difficult than it might seem, Snowflake outperformed Copilot by combining Meta's Rama3 and Mistral-L models with the Snowflake SQL model," he said.

Along with this, Snowflake has released an interface that allows users to create chatbots in minutes with no code, without any specialized knowledge, through 'Snowflake AI & ML Studio'.

He also said that for Snowflake Enterprise AI, “we are continuously investing in improving performance such as LLM lightweightness and accuracy,” and that “Snowflake will be able to satisfy the cost optimization part that enterprises consider most important, such as fine-tuning with the fewest number of parameters.”

On this day, Snowflake also shared updates to its governance model, Snowflake Horizon. To increase open data and interoperability in the AI data cloud, it added features such as Internal Marketplace, Universal Search, Neeva, and AI Powered Object Description.

Snowflake announced new features including: Snowflake Notebooks, Snowpark Pandas API, enhanced DevOps and observability capabilities, Snowflake Trails, and integration of Snowpark Container Service with the Snowflake Native App Framework.