Fivetran announces the launch of the Open Data Infrastructure (ODI) Data Access Benchmark, a public benchmark accessible at opendatainfrastructure.com. This tool measures how major enterprise software vendors facilitate (or restrict) customers’ access to their own data, revealing friction points such as API limits, throttling, and egress charges.
The ODI is an architectural approach promoted by Fivetran to enable organizations to access, move, and use their data without restrictions across systems. Launched as part of a broader initiative, including the merger with dbt Labs in October 2025 and the donation of SQLMesh to the Linux Foundation, it stands in opposition to the closed ecosystems of cloud giants like Databricks or Snowflake.
This benchmark comes at a timely moment: as organizations deploy AI in production, they often discover that their SaaS providers restrict access to the data necessary to feed these workloads.morningstar+2
These factors determine whether the data can be reliably exploited for analytics, operations, and AI. Fivetran emphasizes independence: the scores are objective, with no paid influence, and are updated continuously.
The site opendatainfrastructure.com offers an interactive scorecard, illustrating the gaps between vendors and helping companies assess their AI-ready infrastructure at scale.
Fivetran plans to broaden the benchmark to include more vendors and data sources, expanding the visibility gained through integrations with hundreds of enterprise applications.
Benchmark Evaluation Criteria
The ODI Data Access Benchmark evaluates vendors on three dimensions based on their public documentation:
> Coverage: Programmatic accessibility of the main data objects.
> Performance: Support for high-performance replication, including the Change Data Capture (CDC).
> Egress Charges: Absence of additional fees or restrictive clauses for data replication.
