Snowflake Product Overview
From Analytics and AI to Engineering and Apps—power it all in the AI Data Cloud.
THE AI DATA CLOUD
Streamline your architecture. Smash data silos.
Power your AI, apps and data on a unified platform built for the future.
12,062
Global Customers2
3,098
Marketplace Listings1
features
Get Started
Take the next stepwith Snowflake
Snowflake Product overview
Frequently Asked Questions
All Snowflake products—from Data Engineering to Cortex AI—are integrated into the single AI Data Cloud. There is no data movement or silos; access new features and capabilities directly within your existing Snowflake account without additional installation or infrastructure management.
Through Snowflake Cortex. Cortex gives you instant access to industry-leading Large Language Models (LLMs) and vector search capabilities as a fully managed service. You can build chatbots, search apps, and summarize data securely, keeping your data within the governance boundary of the platform.
Governance is centralized via Snowflake Horizon. Whether you are doing BI analytics, training AI models, or sharing data with partners, Horizon applies a consistent set of compliance, security, and privacy policies across the entire platform.
Snowflake generally operates on a usage-based consumption model. You pay for the compute and storage you consume. While most core capabilities are included, advanced features (like certain AI services or specialized compute clusters) may consume credits at different rates. You do not need to negotiate separate licenses to "turn on" standard platform features.
Most generally available (GA) features are automatically accessible within your Snowflake environment. Features in "Public Preview" may need to be enabled by an admin. You can view the full list of available tools in your account console.
Yes! With Snowpark, data engineers and data scientists can use their preferred languages (Python, Java, Scala) to process data directly inside Snowflake. This allows you to build complex pipelines and ML models without moving data to a separate processing engine.


