Introduction: Simplicity in serverless
We are excited to announce that Serverless Workspaces are now in Public Preview on all clouds. With serverless workspaces, creating new workspaces for your data and AI projects is easier than ever before.
Standing up a new workspace has always been a balancing act. You need to consider cloud networking, where data should live, how to set up compute resources, and how to maintain consistent security across environments. All of this must occur before a single analyst or data scientist can begin their work. The result is delays that slow projects down and add more work for administrators.
Serverless Workspaces change that. Instead of days of planning and configuration, you can create a workspace in seconds. Teams get immediate access to serverless compute, pre-configured storage, and the same governance and controls they already use. That means less overhead and faster starts for new data and AI projects.
What Are Serverless Workspaces?
A Serverless Workspace shifts the heavy lifting of provisioning, monitoring, and maintaining infrastructure to Databricks. When you create a new workspace, serverless compute is ready to run right away. Unity Catalog and storage are automatically set up, backed by Default Storage, so you keep the same governance model while avoiding manual setup.
How Does Serverless Workspace Differ From Classic Workspaces
With classic workspaces, you need to make cloud-level decisions first: how to network compute, where data should land, and how to handle inbound and outbound connectivity. None of that work directly creates value for end users, but it’s required before anyone can log in and build.
With a Serverless Workspace, the basics are already in place. Databricks runs the secure, managed infrastructure for compute and storage. You can still use Unity Catalog, SSO, or identity federation, and your workspace settings, without needing to build and maintain the underlying plumbing.
Serverless Workspace Core Capabilities
Serverless Workspaces include a number of built-in features that make them ready to go from day one:
- Default storage for secure, zero-configuration data access: Each workspace provides fully managed object storage. You can create Unity Catalog-managed catalogs, tables, and volumes without having to bring your own cloud storage, supply storage credentials, or configure external locations upfront. Enterprise features like multi-key protection and no direct access to the object storage make sure that you hold the keys to your data.
- Instantly available serverless compute: Workloads run without provisioning a cluster. Databricks automatically allocates and manages the optimal amount of compute, allowing users to focus on writing code and analyzing data without worrying about cluster operations.
- Access to your existing Unity Catalog data estate: All of your existing governed data is available in the new workspace, where existing permissions allow access.
- Bring your own data when needed: You can connect existing cloud storage through Unity Catalog credentials and external locations, just like you do in classic workspaces. That way, you keep lineage and permissions consistent while avoiding duplicate IAM setups (AWS, GCP, Azure).
- Simplified network security: Instead of building NAT gateways, firewalls, or private link endpoints yourself, you can define serverless egress policies and Serverless Private Link rules (AWS, Azure) that apply to all serverless workloads in the workspace.
- Cost visibility and guardrails: Budget policies let you set attribution tags on workloads and analyze spend in system billing tables (AWS, GCP, Azure). That means clear showback or chargeback without relying on users to tag jobs consistently.
- Agile workspace rollout: Because the infrastructure is managed for you, new workspaces can be created and handed to teams quickly, without waiting for custom networking or runtime upgrades.
Choosing between Serverless and Classic workspaces
Both options remain available. The right choice depends on your requirements:
- Choose serverless when you want the fastest path to a governed environment with minimal configuration and management.
- Choose classic when you need a custom VPC design, strict network patterns, or features that aren’t yet supported by serverless compute. Some organizations also prefer to manage the underlying cloud resources directly.
Serverless Workspace Considerations
There are a few things to keep in mind when planning adoption:
- Region availability: Serverless workspaces are only available in regions that support serverless compute (AWS, GCP, Azure).
- Feature surface: A serverless workspace inherits serverless compute constraints (e.g., Python/SQL focus, unsupported legacy APIs). Validate critical workloads before migrating a workspace wholesale (AWS, GCP, Azure).
- Billing for default storage is not yet enabled on Azure and GCP: During this time, Databricks will not charge for default storage use. Databricks will notify customers 30 days before we enable billing for default storage usage on Azure and GCP.
For further details, see the limitations on Serverless Workspaces (AWS, GCP, Azure) and Default Storage (AWS, GCP, Azure).
Conclusion: Cut the Setup, Keep the Control
For organizations that need to roll out environments quickly or empower new teams without delays, removing the upfront infrastructure build is a real win. You keep the governance and controls that matter, but without the wait.
Serverless Workspaces move Databricks from a “configure first, use later” model to “use now, configure what you need.” Default Storage gives you immediate access to governed data. Egress policies let you centralize outbound control. And Unity Catalog ensures governance remains consistent across both classic and serverless environments.
Ready to experience the simplicity and speed of Serverless Workspaces? Try creating a new serverless workspace in your own account today! And if you don’t have an account, try Express Setup to create one in seconds!
