Dec 16, 2025: Notebooks in Workspaces (Preview)

Snowflake Notebooks in Workspaces is now available in preview. This new notebook experience provides a fully-managed, end-to-end environment for data science and machine learning development on Snowflake data, combining the familiar Jupyter notebook interface with enterprise-grade compute, governance, and collaboration capabilities.

Notebooks in Workspaces runs on a Container Runtime powered by Snowpark Container Services, offering preconfigured containers optimized for AI/ML workloads with access to CPUs and GPUs, parallel data loading, and distributed training APIs for popular ML packages.

Key features

Integration with Workspaces

  • Notebooks are files in Workspaces, enabling easy file management and organization.
  • Git integration provides version control and collaboration across development environments.

Updates to compute and cost management

  • CPU or GPU compute pools match your workload requirements.
  • Shared container service connections reduce start-up time and improve resource utilization.
  • Background kernel persistence ensures uninterrupted execution of long-running processes.
  • Simplified idle time configuration prevents unused compute resources from running indefinitely.
  • Service-level External Access Integration (EAI) management applies to all notebooks in the workspace.

Jupyter compatibility

  • Standard Jupyter magic commands for familiar development experience.
  • Pre-installed data science and machine learning packages.
  • Install additional packages via pip, PyPI, or file upload.

Enhanced editing experience

  • Bidirectional SQL and Python cell referencing for seamless language switching.
  • Interactive datagrid and automated chart builder for data visualization.
  • Enhanced minimap with cell status tracking and table of contents.

For details, see Snowflake Notebooks in Workspaces.