Snowflake ML

What it does

Snowflake ML takes in tabular and other data already stored in Snowflake, plus models trained elsewhere, and produces trained models, feature pipelines, batch predictions and real-time inference endpoints. Work happens in Snowflake Notebooks, through a VS Code extension, or via ML Jobs dispatched from an external IDE.

Teams use it to prepare data, manage features in a Feature Store, train with CPUs or GPUs using preinstalled or open-source packages, track experiments, register models and monitor drift. Its distinct point is that the whole lifecycle runs on the same governed data, with lineage and role-based access control. CoCo can automate parts of this workflow from natural language prompts.

Snowflake ML suits data scientists and ML engineers already working in Snowflake. Using it requires a Snowflake account, and cost is consumption-based, driven by compute, storage and inference. Several capabilities, including some training features, remain in preview, and availability varies by feature.

What it can do

• Build and train models in Snowflake Notebooks on Container Runtime with CPUs or GPUs
• Create and serve ML features with batch and online retrieval in the Feature Store
• Log models from any platform in the Model Registry for batch and real-time inference
• Track experiments and compare trained models in Snowsight
• Monitor model performance, drift and alerts with ML Observability and Explainability
• Trace lineage from source data to features, datasets and models

Works with

• VS Code extension
• Cursor
• Snowflake Notebooks
• Hugging Face
• PyPI

Before you choose

• Costs are consumption-based; plan for compute, storage and inference usage using the credit consumption table.
• Some features are in private or public preview, and some are marked coming soon.
• Models trained outside Snowflake can be imported for inference, but governance features apply to models run on Snowflake data.

Listed under

• Data

Get started

Sign up for a trial and try an ML quickstart.

1. Create a Snowflake account via the free trial sign-up page.
2. Open Snowflake Notebooks and work through the Build Your First Agentic ML Pipeline quickstart.
3. Explore the Snowflake ML developer guide and register a first model in the Model Registry.

Snowflake ML is priced by consumption of compute, storage and inference. The product page links a published credit consumption table for rates.

Source: snowflake.com/en/product/features/end-to-end-ml-workflows, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Snowflake ML | Machine Learning Services and MLOps ↗www.snowflake.com
• Snowflake ML: End-to-End Agentic Machine Learning | Snowflake Documentation ↗docs.snowflake.com
• Getting Started: Quickstarts | Snowflake Documentation ↗docs.snowflake.com
• Snowflake Pricing | Choose the Right Edition for Your Data Needs ↗www.snowflake.com

Company website

snowflake.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.