New industry benchmark uncovers API limits, throttling, and egress fees restrict enterprise data access for AI at scale
Open Data Infrastructure (ODI) is an architectural approach focused on ensuring enterprises can access, move, and use their data across systems without restriction. The ODI Data Access Benchmark is part of this broader initiative, providing a standardized framework to evaluate how data access functions in practice across modern enterprise platforms.
As organizations move AI into production, many are discovering a critical bottleneck: their own software vendors restrict access to the data needed to power AI workloads.
Incomplete APIs limit access to core data objects, throughput restrictions slow replication at scale, and pricing models such as egress fees increase the cost of moving data. These constraints don’t just create friction, they actively limit how quickly and effectively organizations can deploy analytics and AI in production.
Also Read: Tredence Launches Gemini AI Accelerators on Google Cloud
“What we’re seeing is a shift where access to data is becoming more restricted, both technically and commercially,” said George Fraser, CEO of Fivetran. “Oftentimes, those constraints show up as hidden costs or dependencies that push companies toward specific walled-garden ecosystems. As AI becomes more central to how companies operate, organizations need an open data infrastructure that gives them the flexibility to access, move, and use their data without lock-in.”
The ODI Data Access Benchmark evaluates vendors across three dimensions based on publicly documented capabilities:
- Coverage: Whether core data objects are programmatically accessible
- Performance: Whether vendors support high-throughput replication, including change data capture (CDC)
- Egress Charges: Whether data can be replicated without additional fees or restrictive terms
Together, these factors determine whether data can be reliably used for analytics, operations, and AI applications.
The ODI Data Access Benchmark establishes a standard for evaluating how enterprise vendors enable or restrict access to customer data, bringing greater transparency to how data access is structured across the industry. It helps organizations assess whether their data infrastructure can support AI at scale.
Fivetran is calling on software vendors to provide complete, high-performance, and cost-effective data access, and on organizations to evaluate whether their systems can meet the demands of the AI era. The company plans to expand the benchmark over time to include additional vendors and data sources.
Methodology and Transparency
The ODI Scorecard is published by Fivetran, a data movement platform that integrates with many of the vendors evaluated. As a platform operating across hundreds of enterprise applications, Fivetran has direct visibility into how differences in data access models impact customers in practice.
Ratings are determined by Fivetran using publicly available documentation and reflect generally available capabilities only. No vendor has paid for or influenced their score, and the evaluation framework is applied consistently across all vendors, including those Fivetran integrates with commercially. Ratings are reviewed and updated on a rolling basis as new information becomes available.
Source: Businesswire