Cloud data platform whose Cortex AI suite brings managed LLMs, vector search, and agentic workflows directly to governed enterprise data. By 2026, Snowflake Cor
Snowflake with Cortex AI brings models to governed data: SQL-native LLM functions, Cortex Analyst for natural-language analytics, Search for RAG, and agent capabilities, letting analysts and apps use AI where enterprise data already lives, under its access controls.
Pricing: Consumption-based credits across compute and Cortex AI functions; enterprise contracts at scale (published pricing, mid-2026).
Databricks vs Snowflake (2026): Both converged on data+AI platforms from opposite shores. Databricks leans engineering-native: lakehouse, notebooks, training, Mosaic AI agents. Snowflake leans analyst-native: SQL-first Cortex AI, governed simplicity. Existing estate and team skill profile decide more than feature checklists.
LLM functions callable from SQL (summarize, classify, extract), natural-language querying over governed semantic models, and document search: AI capability without leaving the warehouse or its permissions.
Snowflake leans analyst/SQL-native simplicity; Databricks leans engineering/ML depth. Existing platform gravity and team skill profile usually decide: both now cover the data+AI story.