Platform Keynote - New Features

Snowflake Summit 2025 is here! Below is my no-fluff, straight-to-business list of the most exciting features announced at today's Platform Keynote.

Adaptive Compute 

Snowflake automatically determines the appropriate cluster size, number of clusters, and auto-suspend/resume duration for your jobs.

The costly and time-consuming process of determining optimal configurations for compute resources is now (partly) handled automatically by Snowflake. 

Snowpipe Pricing Changes

Snowflake will no longer charge ‘per file’ for Snowpipe ingestions.

This is a huge change for companies that have a high frequency of data being processed through Snowpipe. Snowflake will instead be charging per GB ingested.

Openflow 

Snowflake’s new native data integration tool powered by Apache NIFI.

Users are able to configure all of their data ingestions without requiring a third-party ETL/ELT tool, simplifying pipelines and reducing costs.

Workspaces

A file-based development environment.

Users have the option to utilize a more sophisticated work environment for complex projects that include multiple files.

DBT Projects now on Snowflake

DBT is now available directly inside of Snowflake.

This allows users to build, deploy, and monitor their DBT projects inside of Snowflake with no additional setup or integrations required. 

Snowflake Postgres

A fully managed Postgres database available directly inside of Snowflake.

Existing Postgres projects can integrate with Snowflake without an expensive migration, and new projects can take advantage of Postgres for application development or transactional database needs. 

Snowconvert + AI

Snowconvert is incorporating AI to provide faster and more accurate code translations. 

This feature automatically tests the code that Snowconvert translates to ensure that the logic does not change between the original syntax and the newly generated syntax. 

Gen2 Warehouses

Snowflake’s next generation of warehouses that are more efficient than ever. 

These new warehouses will be slightly more expensive than the existing Gen1 warehouses, but the metrics look promising and will likely be worth transitioning to.

Cortex AI SQL

Cortex has new functions available to use that support text, documents, images, and audio.

A very exciting suite of features that includes aggregate functions and even AI joins, which intelligently determine how to join multiple datasets together. 

Semantic Views

Datasets designed to bridge the gap between how data is stored and what it actually means. 

Semantic views power Snowflake Intelligence and Cortex Analyst, which allows business users to explore their data with natural language instead of SQL queries. 

Snowflake Intelligence

Allows users to intelligently and easily ask natural language questions of their data. 

Snowflake looks through past query history, Tableau dashboards, and third party data such as Stack Overflow or The Washington Post to incorporate any and all available data for natural language interaction with Snowflake data. This brings a completely new level of accuracy to AI responses. 

Inline copilot 

Copilot is available directly inside of worksheets to be used inline while writing queries. 

Using the shortcut CMD/CTRL + i, users can summon a copilot window and ask natural language questions to generate SQL queries. 

Wrapping Up!

Stay tuned for more from Snowflake Summit 2025, including deep dives into several of these exciting features as well as any additional announcements at tomorrow's keynote!