Lakeflow Connect Free Tier is the free usage quota for managed connectors that Databricks announced at FabCon 2026. Each Workspace gets 100 DBU per day, equivalent to about 100 million records ingested for free. That works out to roughly $12,775 (about 1.9 million yen) in annual savings per Workspace — a game-changing move aimed squarely at replacing legacy ETL SaaS like Fivetran and Stitch.
This article walks through exactly what's included in the Free Tier, the full list of supported connectors, a head-to-head comparison with other ETL tools, setup steps, and production best practices.
Lakeflow Connect is Databricks' managed SaaS / DB data ingestion service. Until now, ingesting SaaS data into Databricks required external ETL tools like Fivetran, Stitch, or Airbyte, or hand-rolled Python scripts. Lakeflow Connect handles this through first-party managed connectors from Databricks, replacing all that work with configuration alone.
The Free Tier announced at FabCon 2026 gives Lakeflow Connect 100 DBU of free credits per day per Workspace. The DBU consumed when a connector runs is deducted from this free quota first.
Translating 100 DBU into ingested records works out to up to 100 million records per day. That's more than enough to cover the daily ingestion volume of most mid-sized companies.
Concrete examples:
Premium tier DBU pricing is $0.35. 100 DBU/day × 365 days × $0.35 = $12,775/year, or roughly 1.9 million yen. That's the per-Workspace subsidy.
With a multi-Workspace setup (e.g. dev, staging, prod = 3 Workspaces), you get about 5.7 million yen of value per year. That essentially covers the full ETL tool budget of a mid-sized company.
The following managed connectors are covered by the Free Tier (as of May 2026).
You can cover marketing, sales, HR, finance, IT, and engineering — every department — with the Free Tier alone.
The following are not covered by the Free Tier and incur standard charges.
A dedicated low-latency streaming ingestion feature used for real-time scenarios (seconds-to-minutes latency). Not included in the Free Tier — standard DBU charges apply.
Ingestion from streaming platforms like Apache Kafka uses Structured Streaming and is outside the Free Tier. Organizations with real-time event infrastructure need a separate budget.
Third-party connectors not officially built by Databricks are excluded. Many of these may eventually become official, so consider switching to official connectors based on your requirements.
File ingestion from S3 / ADLS / GCS (Auto Loader) incurs standard DBU charges. Ingestion from an existing data lake remains a paid workload.
Example: ingest Contact data from HubSpot into Databricks on an hourly schedule.
The Workspace admin enables Lakeflow Connect from Settings. Skip if it's already enabled.
From the sidebar, click Lakeflow Connect → New Pipeline → pick HubSpot from the connector list.
Sign in to HubSpot via the OAuth flow and grant read permissions.
Choose the objects you need from Contacts, Deals, Companies, Tickets, Notes, etc. (multi-select supported).
Select the Catalog / Schema. Table names are generated automatically (e.g. `hubspot_contacts`).
Specify the run frequency in cron notation (e.g. hourly = `0 * * * *`).
From here on, ingestion runs automatically. The Unity Catalog tables become available from Delta Lake, DBSQL, and Genie Space.
For data that doesn't need real-time freshness, lower the frequency to minimize DBU consumption — e.g. hourly to every 6 hours. Sum across all your connectors to make sure you stay within the 100 DBU/day quota.
The 100 DBU/day quota is granted per Workspace, so by separating dev, staging, and prod into different Workspaces you effectively get 300 DBU/day worth of free quota.
Build an end-to-end flow: Lakeflow Connect pulls SaaS data → Lakebase handles OLTP processing → Lakehouse Sync writes back into Delta tables. A powerful backend for marketing automation and personalized product experiences.
Right after ingestion, use Lakehouse Monitoring to auto-check data quality. Alert Slack / PagerDuty on anomalies so bad data never flows downstream.
Monitor daily Lakeflow Connect DBU consumption via `system.billing.usage` in System Tables. If you're trending toward exceeding the free quota, lower the frequency or narrow the ingestion scope.
Organizations paying $1,000-5,000 per month for Fivetran or Stitch can cut nearly all of that cost. That's 1-5 million yen per year that can be redirected to actual analytics and ML development.
For startups that can't yet afford ETL tooling, this offers professional-grade ETL for free. Establish a data-driven foundation early.
Organizations using Databricks + Snowflake + BigQuery together risk double-managing tools if they swap part of their stack for Lakeflow Connect alongside a multi-platform ETL tool like Fivetran. Adopt it after committing to a gradual migration or to consolidating on Databricks.
As of May 2026, Lakeflow Connect isn't in the official Exam Guides for Data Engineer Associate or Professional. That said, the following make it very likely to be covered in the next revision:
Recommended topics to study ahead:
How much is included in the Lakeflow Connect Free Tier?
Each Workspace gets 100 DBU per day for free, which translates to roughly 100 million records ingested per day via eligible managed connectors. At Premium tier pricing of $0.35 per DBU, that adds up to about $12,775 USD per year of subsidy per Workspace.
Which data sources are supported?
SaaS apps: Google Analytics, Workday, ServiceNow, Dynamics 365, Zendesk, Google Ads, HubSpot, TikTok Ads, Jira, Confluence, Meta. Databases and DWHs: SQL Server, Oracle, PostgreSQL, MySQL, Snowflake, BigQuery. All of them are covered by the Free Tier, so you can integrate marketing, sales, HR, and finance data across departments at zero cost.
What is excluded from the Free Tier?
(1) Zerobus Ingest (a low-latency streaming ingestion feature) (2) Structured Streaming sources (e.g. Kafka) (3) Community connectors (third-party, non-official) (4) File source connectors (e.g. Auto Loader for S3 / ADLS / GCS). Standard charges still apply to these. The Free Tier only covers managed SaaS and DB connectors.
Will I be charged if I exceed 100 million records per day?
Yes. Usage beyond 100 DBU is billed at the standard Premium rate of $0.35 per DBU. That said, 100 million records per day easily covers the daily ingestion volume of most mid-sized companies, so many organizations can stay within the free quota. For startups and mid-sized businesses, this is a game-changer that can cut ETL costs by hundreds of thousands to millions of yen per year.
Can I ingest data from Snowflake / BigQuery into Databricks?
Yes, both are covered by the Free Tier. This means a multi-platform setup where you store data in Snowflake and analyze it in Databricks becomes possible at zero cost. It also signals Databricks' serious commitment to a multi-platform strategy, and is a tailwind for organizations considering a gradual migration from Snowflake.
When can I start using the Lakeflow Connect Free Tier?
It's already GA on Azure Databricks and is rolling out progressively on AWS and GCP Workspaces. To use it, the Workspace admin just enables Lakeflow Connect and configures the connectors. No additional credit card or separate contract is required — everything is handled within your existing Databricks billing.
How long does setup take?
15-30 minutes per connector is typical. For example, to ingest from HubSpot: (1) click New Pipeline in Lakeflow Connect (2) authenticate via HubSpot OAuth (3) pick the objects to ingest (Contacts / Deals / Companies etc.) (4) specify the destination Catalog / Schema in Unity Catalog (5) set the schedule (e.g. hourly). It takes just a few clicks, no code required.
Can the Free Tier replace Stitch / Fivetran / Airbyte?
Yes. Fivetran charges by MAR (Monthly Active Rows) and runs into the millions to tens of millions of yen per year at mid-sized companies — the Lakeflow Connect Free Tier can drive that close to zero. That said: (1) Fivetran has far more connectors (500+) (2) Airbyte has a fully free OSS edition (3) Stitch / Fivetran are more mature. A practical split: pick Lakeflow Connect if you're consolidating on Databricks, and keep a traditional ETL tool if you're distributing data across multiple platforms.
Start prepping for Databricks exams now
6,800+ practice questions covering all 7 certifications, in real exam format
Try free questions → →Related reading — Databricks new features and adjacent topics
Practice with certification-focused question sets
Test yourself with the Databricks question bankNicheeLab Databricks Editorial Team
NicheeLab editorial team focused on data engineering and cloud certification learning. Content is structured around practical study needs and official exam domains.
Databricks Certifications: All 7 Exams, Difficulty & Study Plan (2026)
Complete guide to all 7 Databricks certifications — Data Eng...
Databricks Exam Difficulty Ranking: All 7 Certs Compared (2026)
Every Databricks certification ranked by difficulty, with st...
Databricks Study Guide: Fastest Pass Route & Time Estimates (2026)
How to pass Databricks certifications efficiently. Official ...
Databricks Data Engineer Associate: Complete Guide (2026)
Domain-by-domain breakdown of the Databricks Certified Data ...
Databricks Data Engineer Professional: Complete Guide (2026)
Tactics for the Databricks Certified Data Engineer Professio...