AI is now a common part of software delivery, but it is creating a security and review bottleneck. According to JetBrains’ State of CI/CD in 2026 report, 48.8% of surveyed organizations use AI for at least one CI/CD-related task, such as reviewing code, writing tests, or diagnosing build failures. However, only 8% run AI-powered steps directly inside a build or test. Teams that do use AI say the sheer volume of generated changes slows down code review and makes it harder to assess reliability — a gap that can let security issues slip through.
Privacy and intellectual property concerns are also holding back adoption. Organizations that report no AI use in CI/CD cite low priority, data privacy, IP worries, satisfaction with existing setups, and cost. Smaller companies are especially likely to say their current approach works well enough without AI. For those that do adopt AI, the report notes additional worries: unpredictable costs, difficulty tracking which changes were made by AI, and debugging problems it introduces.
Self-hosting remains a key response to these concerns. The report finds 72.8% of respondents run their primary CI/CD system on their own premises or on self-managed cloud infrastructure, with on-premises installations reaching their highest share in five years. This reflects teams’ need for control over hardware, networks, data, build environments, security, and costs. As Olga Bedrina of JetBrains put it, self-hosting stays important for organizations that need more control over infrastructure, data, security, or costs — even as AI use around the pipeline grows.