- The dbt engine is free to run. dbt v2 (which replaced Fusion), dbt-oss, and dbt Core (v1) cost nothing to download. The cost shows up in the engineers who build and maintain the platform needed for governed development environments and to run scheduled jobs.
- The dbt platform has four plans, two have published prices. Developer is free and Starter is $100 per seat per month. Enterprise and Enterprise+ are negotiated contracts.
- dbt Labs meters several things. Model builds, Semantic Layer queries, dbt State ($0.094 per daily active target table), and dbt Wizard tokens all land on the consumption bill.
- dbt State can be used even if you self-host or when using a platform like Datacoves. It’s sold as a stand alone add-on that runs against dbt anywhere.
- Datacoves bundles Managed dbt and Managed Airflow in your private cloud or as SaaS, priced per user with no cap on how many users you add, and no per-model metering.
Yes, dbt is free to use. dbt Labs now uses “dbt” to mean dbt v2, its Rust-based engine. It costs nothing to download and run on your own. dbt-oss, the Apache 2.0 subset of dbt v2, and dbt Core, now called dbt v1, are fully open source.
dbt Cloud, renamed to dbt Platform, is a paid managed platform for dbt which includes additional enhancements. It has a free Developer plan for one person, a Starter plan at $100 per seat per month with five developer seats and one project limit, and custom-priced Enterprise and Enterprise+ plans. Usage-based add-ons like dbt State and dbt Wizard can be added on top of any plan. Datacoves also offers Managed dbt and is an alternative to the dbt Platform.
The license is the smallest line in the budget. What enterprise teams pay depends on seat count, metered usage, the end-to-end orchestration when needed, and the engineering time to run anything you don’t buy.
Is dbt Free?
Yes, but “dbt” now means more than one thing. At dbt Summit in September 2026, dbt Labs consolidated its engines into dbt v2 and renamed dbt Core to dbt v1. Here’s what each one costs.
The practical split is simple. If your organization requires Apache 2.0 licensing, run dbt-oss or stay on dbt v1. If it doesn’t, dbt Labs recommends the full dbt v2 distribution, which is free but bundles some proprietary components like connecting to dbt Platform for dbt State functionality.
dbt Labs sells add-ons with no seats and no minimums that run against self-hosted dbt or via a managed dbt platform like Datacoves. So teams that self-host or use a managed dbt platform like Datacoves can use features only available with a dbt Labs commercial agreement.
So yes, dbt is free to run. Running it in production is another story. Open source looks free the way a free puppy looks free: the adoption fee is zero, and the vet bills start next week.
What Happened to dbt Fusion?
dbt Labs launched the Fusion engine in May 2025 as a Rust rewrite of dbt. It was free to use under the Elastic License 2.0, which isn’t an open-source license, and it ran alongside the Python-based dbt Core. Teams had to pick an engine to standardize on, and managed dbt platforms like Datacoves could not distribute it.
dbt v2 ended the split. It became generally available in September 2026 as a single Rust engine built on Fusion’s work, and dbt Labs is phasing out the Fusion name. The full dbt distribution is what Fusion became and it now has a more permissive license allowing managed dbt platforms like Datacoves to distribute it. dbt-oss is the same engine with only the Apache 2.0 components.
Pricing didn’t change. The full distribution is free to use, dbt-oss is free and fully open source, and paid features like dbt State are optional add-ons. If your team has been evaluating Fusion, you’re now evaluating dbt v2. For adapter support, migration steps, and what to do with existing dbt Core projects, see our breakdown of dbt Fusion and dbt 2.0.
Get our free ebook dbt Cloud vs dbt Core
Comparing dbt Core and dbt Cloud? Download our eBook for insights on feature, pricing and total cost. Find the best fit for your business!
Get the PDF
dbt Cloud Pricing: Every Plan Explained
dbt Cloud, now the dbt Platform, prices on developer seats plus metered usage. Here’s what each plan includes as of October 2026, according to the dbt Labs pricing page.
Developer Plan (Free)
One seat, one project, 3,000 successful model builds a month. It’s a sandbox for learning dbt or evaluating the platform. No team can run production on it.
Starter Plan ($100 per Seat per Month)
Starter is the first team plan. Five developers cost $500 a month before usage, billed monthly to a credit card. The limits matter more than the price. Starter gives you five developer seats and one project, and it leaves out SSO. The day you hire a sixth analytics engineer, split work into a second project, or your security team asks for SSO, you’re in an Enterprise plan.
Enterprise and Enterprise+ (Custom Pricing)
Enterprise is where most mid-size and large organizations land, because that’s where teams larger than 5 dbt developers, dbt Mesh, column-level lineage, and SSO, RBAC, and SCIM live. It’s billed annually by invoice, and onboarding and professional services cost extra. Enterprise+ adds unlimited projects, PrivateLink, IP restrictions, and Hybrid Projects for running dbt.
Neither tier has a published price. Your number depends on seat count and committed usage. Budget for a procurement cycle, and get overage and add-on rates in writing before you sign.
Usage-Based Add-Ons: dbt State and dbt Wizard
In 2026, dbt Labs moved two products to pure consumption pricing. Both are optional, and both bill separately from your seats.
dbt State deserves a closer look. It skips, clones, or reuses models that haven’t changed, and dbt Labs reports an average 15 to 30% reduction in warehouse compute. Whether it saves you money depends on how your warehouse bill compares to the DATT charges it creates. Run the math on your own project before you turn it on across production.
Model Build Overages
Every plan includes a monthly allowance of successful model builds. Builds past that allowance are billed as overage. dbt Labs no longer lists the per-model rate on its pricing page. Starter has long been billed at $0.01 per model, and Enterprise customers negotiate their own rates. Ask for the current number before you size your schedule.
Each charge is small on its own. Stacked together, they turn a predictable seat price into a variable bill that can move when someone adds a dbt model or schedules another job.
What a dbt Cloud Bill Looks Like
Take a five-person team on Starter. Their project has 400 models and about 800 tests, and it runs three times a day. That’s 36,000 successful model builds a month. They’ve turned on dbt State, and it reuses roughly 600 target tables (models and tests combined) each day.
- Seats: 5 × $100 = $500
- Model overage (estimated at $0.01): (36,000 − 15,000) × $0.01 = $210
- dbt State: 600 DATT × $0.094 × 30 days = $1,692
- Monthly total: about $2,400, or roughly $28,800 a year
Notice which line is biggest. The seats are the predictable part. dbt State can earn its keep if it cuts enough Snowflake, Databricks, or BigQuery spend, but it’s a third meter that grows with your project. Drop it, and the bill is about $710 a month.
Then comes growth. Hire a sixth developer, add a second domain project, or require SSO, and you’ve left the published price list entirely.
What dbt Cloud Pricing Doesn’t Cover
The dbt platform prices the T in ELT. Enterprise data teams run a lot more than transformation, and the rest of the stack carries its own costs.
- Orchestration beyond dbt. The built-in scheduler runs dbt jobs. Ingestion, Python workloads, ML pipelines, and cross-system dependencies need a full orchestrator, usually Airflow, which you pay for separately.
- Warehouse compute. Every model build runs on your warehouse bill, regardless of which dbt you use.
- Ingestion. Moving data into the warehouse is another tool and another contract.
- Private cloud deployment. The dbt platform is hosted by dbt Labs. Hybrid Projects on Enterprise+ let you run dbt on a dedicated cloud account, and the IDE, scheduler, and catalog stay in dbt Labs’ cloud. For pharma, financial services, and government teams that need the whole platform inside their own account, no plan gets there.
This is why the sticker price may undercount the total bill. You’re buying one layer and assembling the rest yourself. For a feature-by-feature view, see dbt Core vs dbt Cloud.
What It Costs to Run dbt Core Yourself
Skipping the license doesn’t skip the work. Whether you run dbt v1, dbt-oss, or the full dbt v2 distribution, assume two to four engineers earning $130,000 to $170,000 a year spend 25 to 50% of their time building and maintaining a dbt and Airflow platform. That’s $5,500 to $28,000 a month in salary alone, before benefits or cloud infrastructure. At enterprise scale, it turns into a dedicated platform team and an ongoing support and upgrade load for a one-off platform.
Here’s where that time goes:
- Developer environments. Python virtual environments, pinned dbt versions, VS Code configuration, credentials, and onboarding every new hire into the same setup.
- CI/CD. PR validation, SQL linting, test gates, docs generation, and deployment pipelines that someone has to build and keep working.
- Airflow on Kubernetes. Scaling, high availability, DAG syncing from Git, and upgrades without breaking production.
- Security. Secrets management, SSO, audit logging, and the compliance reviews that come with them.
- Upgrades. The move from dbt v1 to v2 is a live example. Adapters for Postgres, Fabric, and Athena are still on the way, and somebody has to test every adapter and package before production moves.
Then there’s the risk nobody budgets for. When the engineer who built the platform leaves, the knowledge leaves with them. Building your own data platform is like wiring your own home: possible, but very few people should. Our build vs buy analysis walks through that decision in detail.
When dbt Cloud May Not Make Sense
The dbt Platform works well for some teams. These are the signals you may have outgrown the platform:
- You are on the Starter plan and your team is past five developers, or security needs SSO, and you need to move to an Enterprise tier.
- Model overages and add-on meters make next quarter’s bill hard to forecast.
- You’re paying for the dbt platform and also running Airflow on the side to orchestrate everything else.
- Security or compliance requires the full platform to run inside your own cloud account.
- Your team prefers to use VS Code
If two or more of these sound familiar, the issue is the pricing structure. Switching plans may not fix it.
How Datacoves Pricing Works
Datacoves is an end-to-end data platform built on Managed dbt and Apache Airflow. It deploys in your private cloud (AWS, Azure, or GCP) or runs as SaaS. Pricing has no project or developer caps, so adding developers, projects, or job runs doesn’t mean migrating to another tier. See Datacoves plans for details.
What’s included:
- In-browser VS Code with dbt, Python, Git, SQLFluff, and any Python library and VS Code extension your team uses.
- Managed Airflow with My Airflow for individual developer testing and Teams Airflow for production.
- Pre-built CI/CD for dbt tests, SQL linting, governance checks, and deployments, with any Git and CI provider.
- Private deployment inside your network, with no VPC peering, your SSO, and secrets stored in your own vault.
- Multi-project support with Datacoves Mesh, which does what dbt Mesh does, plus AI integration with the LLM your organization has approved, billed by your provider.
- Onboarding and architecture guidance from engineers who have run these platforms at enterprise scale.
Standardized environments cut onboarding by roughly 30 hours per developer, and Johnson & Johnson stood up its data stack in weeks. For a full feature comparison, see Datacoves vs dbt Cloud.
dbt Core vs dbt Cloud vs Datacoves: Which Should You Choose?
Choose self-hosted dbt (v1, dbt-oss, or v2) if you have a platform team with spare capacity and a strategic reason to own every layer.
Choose dbt platform Starter if you’re five developers or fewer on one project, you don’t need SSO, and you have no private cloud requirement.
Choose Datacoves if you’re an enterprise running multiple projects, you need Airflow alongside dbt, and you want the platform inside your own cloud with costs you can forecast.
Get a Cost Estimate for Your Team
If you’re budgeting a dbt rollout or staring down an Enterprise renewal, book a free architecture review. We’ll map what your current stack costs, including the engineering time, and show you what Datacoves would look like running inside your own cloud.
