dbt Pricing Guide 2026: Costs & Plans Broken Down

dbt costs nothing for one developer on the free Developer plan, $100 per developer per month on Starter, and a quote on Enterprise. dbt Core, the open-source engine underneath dbt Cloud, is free under the Apache 2.0 licence if you run it yourself.

The seat price is the easy part. Below I go through what each plan meters, what model builds, add-ons and warehouse compute add on top, and what a growing team ends up paying.

What dbt Cloud Actually Costs

dbt's published prices were verified against dbt's pricing page on 2 October 2026.

Developer: Free. One seat, 3,000 successful models built per month, 1 project

Starter: $100 per user/month. Up to five developer seats, 15,000 models/month, 5,000 queried metrics, 1 project. Adds lineage in dbt Catalog, the Semantic Layer and API access

Enterprise: custom pricing. 100,000 models/month, 20,000 queried metrics, 30 projects

Enterprise+: custom pricing. Same model limits, unlimited projects, plus PrivateLink, IP restrictions and hybrid projects

What the tiers actually meter is models built, not seats: 3,000 on Developer, 15,000 on Starter, 100,000 on Enterprise. Seat count is the more visible number and the model count is the one that forces the upgrade.

But the seat price is only part of the bill. Three more lines can land on top of it:

• Successful models built beyond your plan’s monthly allowance
• dbt State and dbt Wizard, two newer add-ons billed by use, if you switch them on
• Your data warehouse compute (Snowflake, BigQuery, Databricks, Redshift), billed by the warehouse rather than by dbt

How dbt Cloud Pricing Works

dbt bills for two things: the seats you hold and what you use each month. The usage rules below come from dbt’s billing documentation, last updated on 18 September 2026.

Developer Seats

What needs a Developer seat: anyone who works in the dbt platform day to day, writing, editing or deploying models.

What doesn’t: dbt also sells Read-Only seats for people who only view documentation and reports, and IT seats for account administration. People who use the transformed data downstream, in a BI tool, need no dbt seat at all.

This distinction matters. Many teams assume they need seats for all data analysts, then realize only the people building in dbt need Developer seats.

Successful Models Built

The usage meter is the number of models dbt builds successfully in a deployment environment each month. That covers scheduled jobs, CI builds triggered by pull requests, and runs started through the API. A job that fails after building 51 of its 100 models counts 51.

What doesn’t count matters as much. Models built in the development environment don’t count, and neither do tests, seeds, snapshots or ephemeral models. Older guides, including an earlier version of this one, said development refreshes ate into the allowance. dbt’s documentation says they don’t.

So the number to estimate is your model count multiplied by how often production runs. A project with 150 models built once a day uses about 4,500 builds a month: over Developer’s 3,000, and well inside Starter’s 15,000. Rebuild the same 150 models every hour and it’s about 108,000 a month, past even Enterprise’s 100,000.

dbt’s pricing page doesn’t list a rate for builds beyond your allowance, so ask for it in writing before you sign.

dbt State

dbt State skips models whose inputs haven’t changed, reusing or cloning them instead of rebuilding. A reused model doesn’t count as a successful model built. It’s billed instead as a daily active target table (DATT): $0.094 for each model, test or snapshot reused in a day, pay-as-you-go, after a 30-day free trial. It works with a dbt Cloud plan or on its own with self-hosted dbt Core.

If State reuses 100 tables every day, that’s 100 × $0.094 × 30 = $282 a month. Whether that saves money depends on what those rebuilds were costing in warehouse compute, which is the bill State exists to shrink.

dbt Wizard

dbt Wizard is dbt’s AI agent for writing and changing the models in your project. It’s billed per token, at each AI model’s own rate. Developer and Starter accounts get a $100 Wizard credit for a 30-day trial and then pay as they go; Enterprise includes $100 a month in credits and Enterprise+ includes $200. You can also bring your own API key and pay the AI provider directly.

Data Warehouse Compute

dbt runs SQL transformations in your warehouse. Every dbt job execution consumes warehouse compute credits that you pay for separately.

Unlike tools that process data in their own infrastructure, dbt pushes all transformation logic to your Snowflake, BigQuery, Databricks, or Redshift environment.

This means your warehouse bill increases with dbt usage, similar to Matillion but without an additional transformation platform license on top.

Real Cost Example: Growing Data Team

Here’s the dbt side of the bill for a team that grows, using only dbt’s published prices.

Year One

Three analytics engineers on Starter: 3 × $100 × 12 = $3,600 a year. The project has 150 models built once a day, about 4,500 successful builds a month, comfortably inside Starter’s 15,000.

Year Two

The team grows to five, Starter’s limit: 5 × $100 × 12 = $6,000 a year. The project grows to 400 models, and the business wants them refreshed every four hours: 400 × 6 × 30 = 72,000 builds a month, nearly five times Starter’s allowance.

At that point the seat count isn’t what forces the upgrade. The build count is, and the next step is an Enterprise quote. A sixth developer would force the same conversation, because Starter stops at five seats.

For a sense of what that quote comes to, dbt Labs’ AWS Marketplace listing sells a 10-developer plan for $48,000 a year, $400 per developer per month and four times Starter’s rate. Vendr, which negotiates software purchases for its customers, reports a median dbt Cloud contract of $27,930 a year across 156 purchases (read on 2 October 2026).

Warehouse compute sits on top of both years, on your Snowflake, BigQuery or Databricks bill. I haven’t put a number on it, because it depends on your data volumes and your warehouse’s pricing, not on dbt.

dbt Core vs dbt Cloud: Cost Comparison

dbt offers two deployment options with dramatically different cost structures:

dbt Core (Open Source)

Cost: $0 (Apache 2.0 license, fully open source)

What you get:

• Complete transformation capabilities
• All modeling features
• Command-line interface
• Local development environment

What you manage:

• Orchestration (using Airflow, Dagster, Prefect, etc.)
• Job scheduling
• Development environment setup
• Version control integration
• CI/CD pipelines
• Documentation hosting
• Model lineage visualization (DIY)

Who it works for: Engineering-heavy teams comfortable with DevOps, organizations with existing orchestration infrastructure, budget-constrained projects.

dbt Cloud

Cost: $100 per developer per month on Starter, up to five developers; Enterprise by quote

What you get (beyond Core):

• Cloud-based IDE
• Job scheduler built-in
• Automatic documentation generation and hosting
• Visual model lineage (DAG)
• Development environments managed for you
• CI/CD integration
• SOC 2 compliance and security features
• Support from dbt Labs

Who it works for: Teams wanting managed infrastructure, organizations prioritizing speed over cost, companies needing compliance features.

The Hidden Costs of dbt Cloud

Beyond the seat price, a few costs catch teams off guard:

CI Builds Count Against Your Allowance

Every pull request that triggers a CI build counts the models it builds. On a 150-model project, ten pull requests a day that each rebuild the whole project add 1,500 builds a day. Setting CI to build only the models a change touches keeps that number small.

Model Complexity and Warehouse Costs

Complex dbt models with multiple CTEs, window functions, and large joins consume significant warehouse compute, and every scheduled rebuild pays for it again. That repeated cost is the one dbt State is designed to cut.

Support and Services

dbt’s pricing page lists onboarding and professional services as something Enterprise customers can buy, without a published rate.

Training Costs

dbt assumes SQL and Git. For an analyst who already writes SQL, the learning is dbt’s conventions. For one who doesn’t, it’s SQL first, and that’s the larger cost.

How dbt Cloud Compares to Alternatives

Compared on what each one needs from you, not only on price:

Matillion and Informatica bill in ways that don’t reduce to one entry figure, so I’ve left them out of the table; Matillion pricing and Informatica pricing go through how each one charges. One more change worth knowing: Fivetran, the loading tool most often paired with dbt, has merged with dbt Labs, so the two are now one vendor.

When dbt Cloud Makes Sense

dbt Cloud excels in specific scenarios:

You’re building a modern data stack. If you’re already committed to Snowflake, BigQuery, Databricks, or Redshift, dbt Cloud fits naturally into that architecture.

Your team thinks in SQL. dbt is SQL-first. If your analytics engineers write SQL daily, they’ll be productive in dbt within days. The learning curve is about dbt conventions, not learning a new transformation paradigm.

You want transformation logic as code. dbt brings software engineering practices (version control, testing, documentation) to data transformation. For teams valuing these practices, dbt delivers.

Managed infrastructure matters. dbt Cloud handles deployment, scheduling, and environment management. If you’d rather not build and maintain orchestration infrastructure, the cost is worth it.

When to Consider Alternatives

Several scenarios suggest looking elsewhere, and I compare the tools for each in dbt alternatives:

Limited SQL Expertise

dbt requires SQL proficiency. If your team works primarily in spreadsheets or visual tools, the SQL-first approach creates friction.

For teams needing business user-friendly data preparation, visual platforms work better.

Cost Sensitivity

Seats, builds beyond your allowance, add-ons and warehouse compute arrive as separate lines from at least two vendors, which makes the total harder to forecast than any one of them.

If budget constraints are tight, either dbt Core (free but requires engineering) or fixed-cost platforms provide more predictability.

Complex Orchestration Needs

dbt Cloud’s built-in scheduler handles straightforward dependencies. For complex workflows spanning multiple tools, external orchestration (Airflow, Prefect, Dagster) provides more flexibility.

Many teams end up using dbt Core with external orchestration anyway, making dbt Cloud’s scheduler less valuable.

Non-Warehouse Data Sources

dbt works best when all your data is already in your warehouse. If you need to process data from APIs, files, or databases not in your warehouse, you need additional tools.

Getting dbt Cloud Pricing

To get actual pricing from dbt Labs:

1. Sign up for free trial at getdbt.com

The Developer plan is genuinely free and gives you real experience with the platform.

2. Estimate your needs:

• Number of analytics engineers writing dbt models
• How often you’ll run transformations (hourly, daily, weekly)
• Size of your data warehouse
• Whether you need SOC 2, HIPAA, or enterprise security

3. Count your model builds: During the trial, watch successful models built on each project’s home page, which is where dbt reports them. Multiply by how often production will run.

4. Calculate warehouse costs: Track Snowflake/BigQuery compute during trial runs. Project monthly costs based on production frequency.

5. Request enterprise pricing: Contact dbt Labs sales for Enterprise plan details if you need:

• More than five developer seats (Starter’s limit)
• Advanced security features
• Custom SLAs
• Dedicated support

Questions to Ask

About usage: “What do we pay per successful model built beyond our plan’s allowance?”

“If we turn on dbt State, which of our models would be reused rather than rebuilt, and what would that cost in DATTs?”

About scaling: “How do costs change as we add models and increase refresh frequency?”

“Can you show examples of customers with similar usage patterns?”

About warehouse impact: “What’s typical warehouse compute consumption for dbt workloads at our scale?”

About alternatives: “For our use case, do you recommend Cloud or Core? Why?”

dbt Core vs dbt Cloud: Making the Choice

The Core vs Cloud decision dominates dbt pricing discussions:

Choose dbt Core if:

You have data engineering resources. If your team can manage orchestration, CI/CD, and infrastructure, Core saves significant money.

You already run Airflow or similar orchestration. Adding dbt Core to existing orchestration is straightforward.

Budget is extremely tight. Free is unbeatable if you can handle the operational overhead.

You need maximum flexibility. Core gives you complete control over deployment and execution.

Choose dbt Cloud if:

You want to move fast. Cloud gets you productive in hours vs weeks of infrastructure setup.

Your team is analytics-focused, not engineering-focused. Managed infrastructure means analysts can focus on models, not DevOps.

Compliance matters. SSO, role-based access control and SCIM provisioning come with dbt’s Enterprise plan.

Support is valuable. Having dbt Labs support for questions and issues justifies the cost.

The Strategic Calculation

Choosing dbt involves more than pricing. Consider:

Team Skill Set

dbt requires SQL proficiency and software engineering practices (Git, testing, documentation). If your team has these skills, dbt works well. If not, training costs and productivity losses matter.

Modern Data Stack Commitment

dbt assumes you’re building on Snowflake, BigQuery, Databricks, or Redshift. If you’re committed to this architecture, dbt is the natural transformation layer.

If you’re uncertain about your warehouse choice or work with diverse data sources, more flexible platforms might fit better.

Long-Term Warehouse Costs

dbt’s warehouse-native approach means your warehouse bill grows with transformation complexity, and that bill comes from your warehouse vendor, not from dbt.

Budget for warehouse compute as part of total dbt cost. Track it separately to understand true TCO.

Making the Decision

The right transformation tool depends on your specific situation.

Choose dbt Cloud if you’re building a modern data stack, have SQL-fluent analysts, value managed infrastructure, and can absorb combined platform + warehouse costs.

Choose dbt Core if you have data engineering resources, existing orchestration infrastructure, tight budget constraints, and need maximum flexibility.

Consider alternatives if your team lacks SQL skills, needs visual transformation tools, works with non-warehouse data sources, or requires fixed-cost predictability.

Testing Before Committing

dbt Cloud: Sign up for free Developer plan immediately

dbt Core: Download from GitHub and test in your environment

Mammoth Analytics: 14-day free trial for no-code alternative

Fivetran: Trial for automated data ingestion

Testing reveals whether SQL-first transformation fits your team’s workflow and whether warehouse compute costs are acceptable.

The Real Question

dbt Cloud pricing ultimately matters less than whether SQL-based transformation-as-code fits your team’s capabilities and workflow.

Can your team write and maintain SQL transformations effectively?
Does transformation-as-code align with your data strategy?
Are combined platform + warehouse costs acceptable?
Will dbt scale with your growing transformation needs?

Answer these honestly before committing. Many teams succeed with dbt because it matches their skills and architecture. Others discover after months that their team needed more visual, code-free tools.

The best tool isn’t the cheapest or most popular. It’s the one your team will actually use effectively.

Frequently Asked Questions

How much does dbt cost?
dbt Cloud is free on the Developer plan for one developer and 3,000 successful model builds a month. Starter is $100 per developer per month, with five developer seats and 15,000 model builds a month. Enterprise is priced by quote, and dbt’s own AWS Marketplace listing sells a 10-developer plan for $48,000 a year. The warehouse compute your models use is billed separately, by Snowflake, BigQuery, Databricks or whichever warehouse runs them.

Is there a free version of dbt?
Yes, two. dbt Core is open source under the Apache 2.0 licence and free to run on your own machines, with your own scheduler. dbt Cloud’s Developer plan is free for one developer seat and 3,000 successful model builds a month.

How much does dbt State cost?
dbt State costs $0.094 per daily active target table, pay-as-you-go, after a 30-day free trial. You pay for each model, test or snapshot that State reuses instead of rebuilding in a day, so 100 reused tables every day comes to about $282 a month. It works with a dbt Cloud plan or with self-hosted dbt Core.

How much does dbt Wizard cost?
dbt Wizard, dbt’s AI agent, is billed per token at each AI model’s own rate. Developer and Starter accounts get a $100 credit for a 30-day trial, Enterprise includes $100 a month in credits and Enterprise+ includes $200, and you can bring your own API key instead.

What are the downsides of dbt?
dbt expects everyone who builds models to write SQL, and it only transforms data that is already in your warehouse. Loading the data is a separate tool, and every model run is warehouse compute you pay for. If you use dbt Core rather than dbt Cloud, scheduling, hosting and CI are yours to set up too.

---

Related Reading

Explore more about data transformation and modern data stack tools:

• Best ETL Tools for 2025 – Compare platforms
• Matillion Pricing Breakdown – Similar warehouse-native approach
• Data Transformation Tools Compared – Visual vs SQL-based
• What is a Data Pipeline? – Architecture fundamentals
• SQL vs No-Code Data Tools – When to use which approach