dbt on Snowflake vs. dbt Cloud Starter: A Detailed Analysis

Overview

Following the announcement at Snowflake Summit in June, 2025, dbt (data build tool) projects are now natively deployed within Snowflake. With dbt being the go-to framework for analytics engineers, we, like many others, have been asking ourselves: Should you run dbt Core on Snowflake, or invest in dbt Cloud?

We recently conducted a thorough analysis of dbt core on Snowflake vs. dbt Cloud Starter for one of our SMB clients. Thus, the purpose of this article is to share our key insights broken down into the following sections:

1. Development Environment
2. Orchestration & Scheduling
3. CI/CD & Pull Request Workflow
4. Team Collaboration & Scaling
5. Restrictions & Limitations
6. Final recommendations

Development Environment

dbt core on Snowflake requires manual scheduling through Snowflake or external tools, while Starter includes a built-in scheduler — meaning your data pipelines can run reliably without extra setup. This choice affects how quickly and consistently fresh data reaches your team.

• dbt Core on Snowflake – A CLI-only setup means developers write and run dbt commands locally. Ideal for power users comfortable with terminal workflows but creates a steeper learning curve for analysts.
• dbt Cloud Starter – A browser-based IDE, enabling real-time development and code review without local setup.

Takeaway: Organizations prioritizing ease of use and fast onboarding should lean toward dbt Cloud Starter for its intuitive interface.

Orchestration & Scheduling

dbt core on Snowflake requires manual scheduling through Snowflake or external tools, while Cloud Starter includes a built-in scheduler — meaning your data pipelines can run reliably without extra setup.

• dbt Core on Snowflake – Requires manual orchestration through Snowflake Tasks or third-party schedulers (Orchestra, Airflow, Dagster)
• dbt Cloud Starter – Comes with a built-in job scheduler featuring logs, notifications, and retry capabilities — no extra infrastructure required.

Takeaway: If you want automated scheduling for dbt transformations without maintaining separate tools, dbt Cloud Starter is the better choice.

CI/CD & Pull Request Workflow

dbt core on Snowflake makes you set up testing and pull request workflows yourself, while Starter has GitHub/GitLab integration built in. This difference can make your deployments either smooth and fast or slow and manual.

• dbt Core on Snowflake – CI/CD pipelines must be configured manually. Git PR Workflow will be done the same as it is in the repo and not on dbt cloud. CI/CD is not available​ — a hurdle for teams without strong DevOps support.
• dbt Cloud Starter – Offers native GitHub and GitLab pull request workflows out of the box. While Starter doesn’t include full CI/CD automation (available in Enterprise), it still reduces setup time dramatically.

Takeaway: For dbt PR workflows and smoother Git integration, Cloud Starter shortens the deployment cycle.

Team Collaboration & Scaling

dbt Core on Snowflake works best for small, technical teams used to local dev workflows, while Starter is designed for scaling — with shared environments and centralized control. This choice will influence how easily your analytics engineering team can grow.

• dbt Core on Snowflake – Best suited for small, technically proficient teams where local development and Git discipline are standard.
• dbt Cloud Starter – Designed for scalable analytics engineering, with centralized environments and easier onboarding for distributed teams.

Takeaway: For collaborative analytics engineering at scale, dbt Cloud Starter offers a stronger foundation.

Restrictions & Limitations

There are a few restrictions and limitations to be aware of when deciding between dbt core on Snowflake and dbt Cloud Starter.

Data Governance & Observability

• dbt Core on Snowflake – Offers basic functionality but lacks enterprise-grade controls. It relies heavily on Snowflake’s native RBAC (Role-Based Access Control), which may require additional configuration for full observability.
• dbt Cloud Starter – Provides rich metadata, RBAC, and API access out of the box, making it a better fit for teams that need robust governance and auditability.

Takeaway: dbt Cloud Starter is ideal for organizations prioritizing data governance and compliance.

File Limitations

• dbt Core on Snowflake – Projects are limited to 20,000 files, including logs and packages. This can be a bottleneck for large-scale deployments.
• dbt Cloud Starter – No file limitations, offering greater flexibility for complex projects.

Takeaway: For large or modular dbt projects, Cloud Starter removes structural constraints. Though, log storage can be moved to Snowflake internal stage, S3, or stored in Snowflake tables to workaround the limitations for dbt core on Snowflake.

Execution Limits

• dbt Core on Snowflake – No hard limits on model execution.
• dbt Cloud Starter – Capped at 15,000 successful models built per month.

Takeaway: dbt Core offers more freedom for high-frequency model builds, but Cloud Teams may suffice for most mid-sized teams.

Git Repository Size

• dbt Core on Snowflake – Git repos larger than 2GB are not supported.
• dbt Cloud Starter – No size restrictions, making it more scalable for version control.

Takeaway: Cloud Starter is better suited for teams managing large codebases.

Semantic Layer Capabilities

• dbt Core on Snowflake – Not available.
• dbt Cloud Starter – Supports up to 5,000 queried metrics per month via the dbt Semantic Layer, enabling centralized metric definitions and business logic.

Takeaway: dbt Cloud Starter unlocks advanced analytics capabilities through its semantic layer.

Final Recommendation: Core vs. Cloud Starter

Choose dbt Core on Snowflake if you have a small, technical team that values full control over infrastructure and can manage orchestration independently.

Choose dbt Cloud Starter if you want faster deployment, built-in orchestration, centralized collaboration, and hosted documentation.

Interested in learning more? Connect with an dbt expert today.