Kumo can be deployed in three ways: (a) the Kumo SaaS deployment provides an Apache Spark-based data platform, expanded choice of supported data warehouses, earlier feature access, quicker bug fixes, and easier access to Kumo’s enterprise support tier. (b) The Data Warehouse Native option allows you to bring your own data platform (Snowflake or Databricks) and access Kumo’s highly accurate predictions without raw data stored or materialized outside of your boundary. (c) The Virtual Private Cloud option runs Kumo as a self-contained Kubernetes deployment inside your VPC/VNet with your compute, storage, network, identity provider, and data source permissions. See Virtual Private Cloud for details. For proof-of-concept evaluations, Kumo also offers a Simplified VPC Deployment: a self-contained machine image that can be installed quickly in AWS, Azure, or Google Cloud with no external network egress and Parquet-only data access. Data storage is always in your environment regardless of deployment option, and you can choose the deployment option that best aligns to their organizational data governance policies and preferences. The following table provides an overview of the different deployment options:Documentation Index
Fetch the complete documentation index at: https://kumo.ai/docs/llms.txt
Use this file to discover all available pages before exploring further.
| SaaS | Virtual Private Cloud | Databricks | Snowflake Native App | |
|---|---|---|---|---|
| Customer data warehouse | Amazon Redshift, AWS S3, Google Cloud BigQuery, Databricks, Snowflake | Snowflake, Databricks, BigQuery, S3, other lake/warehouse sources | Databricks | Snowflake |
| Data storage | Your own | Your own object storage and KMS | Your own | Snowflake objects (e.g., tables, views, etc.) in your own account |
| Data cache | Kumo-owned and managed | Customer storage in your VPC/VNet (bucket/blob) | Customer’s Databricks Unity Catalog Volume storage | Snowflake stage in your own account |
| Data platform | Kumo-managed Apache Spark | Customer-owned clusters (EKS/AKS/GKE) connecting to your warehouses/lakes | Databricks Spark in your own Databricks account | Snowpark DataFrame in your own Snowflake account via Snowpark Container Services |
| ML compute | Kumo-owned and managed compute | Customer-managed autoscaling GPU and high-memory nodes inside your VPC/VNet | Kumo-owned and managed compute | Kumo services in Snowpark containers |
| PII and sensitive data handling (GDPR compliance) | Your own retention policy | Kumo services and artifacts run in your VPC/VNet; source access is governed by customer-owned permissions | No data on disk in a Kumo-owned environment; data leaving your environment is transformed and encoded | No data on disk in a Kumo owned environment; no data leaves your environment |
| Access to private preview features | Yes | Yes | Yes | No |
| Access to Enterprise support tier | Yes | Yes (via your VDI/bastion; no persistent Kumo access required) | Yes | Kumo requires temporary elevated access and logs |