Validation at scale becomes a key factor for top teams that more than double throughput year-over-year
CircleCI, the leading software delivery platform for teams to ship confidently at AI speed, today unveiled its 2026 State of Software Delivery report. Based on the analysis of more than 28 million CI workflows across thousands of teams, the seventh annual report, sponsored this year by Thoughtworks, shows the largest increase in development activity since its first publication in 2019.
The data reveals a key theme: Not all teams can accelerate software delivery by simply implementing AI in development practices. There is a widening gap between organizations that can effectively transform AI-driven acceleration into measurable business results and those that cannot. The data shows that success in the AI era is not determined by how quickly code can be written, but by the ability to validate, integrate, and ship code, and recover at scale.
AI isn’t creating a performance gap; it’s revealing it
Across all projects building on CircleCI, average throughput (the number of daily workflows run) grew 59% year over year. AI-powered code generation and agent-driven workflows are enabling teams to explore ideas, generate features, and iterate at unprecedented speed. A closer look at the throughput data shows that the across-the-board, year-over-year improvements are skewed by top-performing teams.
The top 5% of teams nearly doubled their throughput year over year, from 6.8 to 13.4 daily workflow runs. The top 10% and 25% of teams saw smaller but still significant increases of 47% and 25%, respectively. The median team, meanwhile, increased throughput by just 4%, while the bottom 25% saw no measurable increase.
“The industry is far past thinking about how to use AI to accelerate software delivery,” said Rob Zuber, CTO of CircleCI. “The teams that are realizing AI’s benefits in software development are using autonomous validation to run laps around the competition that can’t validate AI-generated code at scale.”
Validation at scale separates winning teams from the rest
Elite teams are now shipping at levels that would overwhelm traditional delivery pipelines. At this scale, the bottleneck is no longer the speed at which developers work. When teams are validating tens of thousands of changes per day, delivery success depends on whether the underlying validation and deployment systems can keep up.
The winning teams are the ones converting raw coding speed into deployable changes. They can translate increased development activity into actual shipped changes. Predominantly, these teams are reducing feedback and recovery cycles to ease the strain caused by higher change volume.
To learn more about how your team stacks up, and to accelerate software delivery by leveraging AI-assisted development:
- Download the full 2026 State of Software Delivery report
- Join us for live discussions around validation in the AI era
Methodology
The 2026 State of Software Delivery Report analyzes data from nearly 28 million workflows from over 22,000 organizations in 149 countries. The data was collected within the first 28 days of September 2025.