Published November 2, 2026
Data integration software copies data from source systems, such as databases, files and SaaS applications, into a central destination for analysis. ETL (extract, transform, load) transforms the data before it is loaded. ELT (extract, load, transform) loads raw data first and transforms it inside the warehouse or lakehouse.
Where the market stands
Salesforce completed its acquisition of Informatica on November 18, 2025. Salesforce says the deal brings Informatica's catalog, integration, governance, quality, privacy, metadata management and master data management services to the Salesforce platform.
Fivetran completed its merger with dbt Labs on June 1, 2026. The all-stock transaction was announced on October 13, 2025. George Fraser remains CEO and Tristan Handy is President. With the merger, the company released an alpha of dbt Core v2.0, which open-sources the dbt Fusion engine runtime under the Apache 2.0 license.
Fivetran signed an agreement to acquire Census, a reverse ETL vendor, on May 1, 2025. Fivetran says Census lets it move data from the warehouse back into business applications. Census joined Fivetran in 2025 and is now called Fivetran Activations.
Boomi acquired Rivery in December 2024. The product is now Boomi Data Integration and adds ELT pipelines and change data capture (CDC) replication to Boomi's integration platform.
IBM completed its acquisition of Confluent, the data streaming company, on March 17, 2026, at $31 per share in cash.
End-of-support dates
- Informatica PowerCenter 10.5.x reached end of support on March 31, 2026. Informatica's path is its Cloud Data Integration service on IDMC. Informatica says customers can import and run PowerCenter workflows as they are, through a Secure Agent, without rebuilding mappings.
- Talend Open Studio, the free open-source edition, was retired on January 31, 2024. Qlik no longer hosts or updates it.
- Apache Airflow 2 reached end of life on April 22, 2026. Airflow 3 was first released on April 22, 2025.
- Azure Data Factory continues, but Microsoft calls Data Factory in Microsoft Fabric its next generation and publishes an upgrade guide for Azure Data Factory and Azure Synapse pipelines.
The main approaches
Products differ in what they move and where the transformation runs. If you need several of these motions, plan for more than one product.
- Managed ELT. The vendor builds, runs and maintains the connectors that extract from sources and load the destination, and transformation follows the load. Fivetran works this way, and Airbyte sells a fully managed cloud service.
- In-warehouse transformation. dbt turns SQL select statements or Python DataFrames into tables and views in the warehouse, with data quality tests written alongside. Microsoft Fabric runs dbt jobs on its own data warehouse.
- Integration suites. Informatica sells integration, data quality, governance, catalog and master data management services on IDMC. In Qlik Talend Cloud, data quality, stewardship and column-level lineage start at the Premium edition.
- Cloud-native services. AWS Glue is a serverless data integration service that runs ETL jobs on Apache Spark. AWS zero-ETL integrations replicate Aurora, RDS, DynamoDB, self-managed databases and applications such as Salesforce, SAP and ServiceNow into Amazon Redshift without a pipeline to maintain. Data Factory in Microsoft Fabric connects to more than 170 sources and supports both ETL and ELT.
- Log-based CDC. Qlik Replicate scans database transaction logs after an initial load. Fivetran sells HVR, a self-hosted version of Fivetran for database replication. Debezium is a set of open-source source connectors for Apache Kafka Connect.
- Integration platforms (iPaaS). Boomi approaches data movement from application integration, with ELT and CDC added through Rivery.
- Orchestration. Apache Airflow is an open-source platform for developing, scheduling and monitoring workflows. Airflow workflows are defined in Python.
How each is priced
List prices as published, read on October 17, 2026.
- Fivetran. Billed on monthly active rows (MAR), on a separate cost curve for each connection. Inserts and updates, including deletes, count. Unchanged rows and the initial sync do not. The Free plan covers 500,000 MAR. Each standard connection with 1 to 1 million MAR carries a $5 base charge. Transformations are billed per successful model run.
- dbt. The Developer plan is free for one seat. Starter costs $100 per user per month and includes five developer seats and 15,000 successful model builds a month. Enterprise pricing is not published.
- Airbyte. Standard starts at $20 a month with 5 credits, and extra credits cost $5 each. Plus starts at $189 a month for 40 credits. Pro and Enterprise Flex are priced on capacity, in Data Workers. Airbyte Core is self-hosted. It is licensed under the Elastic License 2.0, which forbids offering it to others as a managed service.
- AWS Glue. ETL jobs are billed per DPU-hour, by the second. AWS's pricing example uses $0.44 per DPU-hour, or $0.29 with Flex, and says pricing can vary by region. AWS charges no additional fee for zero-ETL integrations, but Glue charges for ingesting application data through them.
- Azure Data Factory. Microsoft's retail price list for East US sets orchestration at $1 per 1,000 activity runs, data movement at $0.25 per DIU-hour and general-purpose data flows at $0.27325 per vCore-hour, with a minimum cluster of 8 vCores. Fabric is sold as capacities sized in capacity units, from F2 with 2 capacity units. Microsoft says a 1- or 3-year reservation saves about 41% against pay-as-you-go.
- Informatica. IDMC is consumption-priced in Informatica Processing Units (IPUs), which give access to all eligible cloud services. No list prices are published.
- Qlik Talend Cloud. Four editions: Starter, Standard, Premium and Enterprise. Usage is measured by data volume moved, job executions and execution duration, against a fixed capacity. No list prices are published.
- Matillion. Developer, Teams and Scale editions, paid in Matillion credits. No per-credit price is published.
What a change requires
With ELT, transformation runs on the warehouse, so its compute appears on the warehouse bill, not the integration vendor's. Count both.
Under row-based pricing, cost follows how often source rows change, not how large the tables are. Model the bill on your highest-change tables at production volume before signing.
A move off PowerCenter, Talend Open Studio or Airflow 2 is a migration project. Inventory every job, its schedule, its dependencies and the systems that read its output. Then pilot on the most complex jobs, not the simplest.
Plan the order of the rollout by business domain. Deliver one domain end to end, from ingestion to a report someone uses, before adding breadth.
Questions to settle first
Settle these before shortlisting.
- Which motions do you need: batch ELT, log-based CDC, reverse ETL, application integration? Which one product covers them, and which need a second tool?
- Where will transformation run, and does your team work in SQL, Python or a visual designer?
- What will the pricing unit cost on your highest-change tables, and who pays the warehouse compute?
- Which jobs run on a product that is out of support, and what is the date to be off it?
- Under its new owner, which warehouses and clouds will your vendor commit to supporting in writing?