Certified Solution
Build Autonomous Pipelines for AI
AITransformation
Snowflake Dynamic Tables
Define the desired state of your data and Snowflake manages the incremental updates and orchestration — no manual coding, no third-party tools.
Overview
Build, orchestrate and optimize
Dynamic Tables handle the hard parts of data pipelines so your team can focus on insights.
Easy to Implement & Test
Build and troubleshoot pipelines with Cortex Code using natural language, or define transformations in a single SQL statement. Native DAG observability — no third-party tools needed.
Simple to Orchestrate
Set data freshness with TARGET_LAG. Snowflake auto-infers dependencies and schedules refreshes to meet your target — no manual DAGs or orchestrators.
Cost-Effective to Operate
Incremental refreshes process only changed data, avoiding full-table recomputation. Lock stable data with immutability and let automatic query mode optimize for your SLAs.
TRUSTED BY DATA TEAMS AT
Key Features
Powerful capabilities built in
Everything you need to build autonomous data pipelines.
Cortex Code
Data engineering coding agent with skills and workflow to build, troubleshoot, and optimize Dynamic Tables using natural language. Available in Snowflake Workspaces and via CLI.
Cortex AI Functions in Dynamic Tables
AI agent that builds, troubleshoots, and optimizes Dynamic Tables using natural language. Available in Snowflake Workspaces and via CLI.
Cost Controls with Immutability
Lock portions of your Dynamic Tables so they never recompute during refreshes. Less recomputation, significantly lower refresh costs.
Dynamic Apache Iceberg™ Tables
Powerful autonomous pipelines now available for your Iceberg Lakehouse.
Custom Incrementalization
Fine-grained control over how incremental refreshes process your data — details coming soon.
Adaptive Refresh Mode
Intelligent refresh mode selection that automatically optimizes for your workload — details coming soon.
USE CASES
Built for your data pipeline needs
From analytics pipelines to real-time change data processing and Lakehouse workflows.
Analytics Pipelines
Transform and model data for analytics, no third-party transformations or orchestrators needed.
- Accessible to anyone who can write SQL
- Control costs with immutability
- Agentic support for authoring and troubleshooting
Change data processing
Continuously transform ingested change data into serving-ready datasets that stay synchronized as upstream data evolves.
Automated incremental updates keep data fresh in under a minute.
Lakehouse Pipelines
Simplify pipeline development within your Open Lakehouse.
- Accessible pipelines defined by a SQL query
- Autonomous management of orchestration and incremental refresh schedule
- Free up development time while delivering stable and reliable transformation pipelines
Get Started
Take the next step with Snowflake
Start your introductory trial with $40 in free credits²
- Accelerate end-to-end development
- Context-aware by default
- Open, extensible and production-ready
DYNAMIC TABLES
Frequently Asked Questions
What is a Dynamic Table in Snowflake and how does it simplify data pipelines?
Dynamic Tables are a declarative way to build data pipelines in Snowflake. Instead of writing and scheduling transformation logic manually, you define the result you want with a SQL query and Snowflake handles the refresh automatically — eliminating the need to orchestrate the pipeline.
How do Snowflake Dynamic Tables differ from traditional ETL pipelines?
Traditional ETL pipelines require custom orchestration, scheduling, and incremental logic. Dynamic Tables replace that complexity with a single SQL definition. Snowflake determines what data has changed and refreshes only what's needed.
How do Dynamic Tables automatically refresh and keep data up to date?
Dynamic Tables use a user-defined target lag — the maximum acceptable time between a source data change and its reflection in the table. Snowflake continuously monitors upstream data and triggers incremental refreshes automatically to stay within that lag window.
What is target lag in Dynamic Tables and how does it affect pipeline latency?
Target lag defines how fresh your data needs to be — for example, TARGET_LAG = '1 minute' means Snowflake will refresh the table within one minute of upstream changes. You can also set DOWNSTREAM to let Snowflake optimize lag based on dependent objects.
How do you monitor and troubleshoot a Dynamic Table pipeline in Snowflake?
Use SHOW DYNAMIC TABLES to check status and lag, and query DYNAMIC_TABLE_REFRESH_HISTORY() in the Information Schema for refresh logs. Snowsight also provides a visual DAG view of your pipeline, making it easy to identify failures, bottlenecks, or tables stuck in an UPSTREAM_FAILED state.
How do Snowflake Dynamic Tables compare to dbt for building data pipelines?
dbt and Dynamic Tables are complementary. dbt excels at transformation modeling, testing, documentation, and lineage — while Dynamic Tables provide the automated, incremental refresh engine that keeps data current without manual scheduling.
With dbt version 1.12+, you can now use Dynamic Tables as a native materialization type directly within your dbt project. This means you get the best of both worlds: dbt's development workflow, testing, and documentation layered on top of Snowflake's automated incremental processing. Simply set materialized='dynamic_table' in your model config along with your desired target_lag, and dbt will create and manage the Dynamic Table for you. It’s easy to test the performance.
For teams already using dbt, this makes Dynamic Tables the recommended path for incremental pipelines on Snowflake.
