JetBrains Report Reveals Shifting CI/CD Landscape: Self-Hosting Rises, AI Integration Remains Peripheral

JetBrains' latest "State of CI/CD in 2026" report, compiled from multiple surveys, reveals several key shifts in how organizations are approaching continuous integration and continuous delivery. A striking finding is the resurgence of self-hosted CI/CD, with 72.8% of organizations now running their primary CI/CD system on-premises or on self-managed cloud infrastructure. This includes a five-year high of 30.5% for purely on-premises installations. The report also highlights that while nearly half of organizations (48.8%) use AI for at least one CI/CD task, only a small fraction (8%) integrate AI-powered steps directly into their build or test processes. This trend towards self-hosting is significant because it indicates a potential re-evaluation of the trade-offs between convenience and control. For many years, the industry narrative has pushed towards fully managed cloud services for CI/CD, promising reduced operational overhead. However, the data suggests that organizations are increasingly prioritizing factors like data sovereignty, security posture, cost predictability, and the ability to deeply customize their environments. This could be driven by a maturing understanding of cloud costs, stricter compliance requirements, or a desire to leverage existing on-premises investments more effectively. The limited in-pipeline AI adoption, despite broader AI usage in development workflows, points to ongoing challenges in making AI reliably and practically contribute to core CI/CD execution, such as automated code generation or intelligent test selection. This development fits within a broader trend of platform engineering teams seeking greater ownership and customization of their development toolchains. While cloud-native principles remain dominant, the implementation is evolving to include more hybrid and self-managed components, especially for critical infrastructure like CI/CD. The rise of GitOps, for instance, emphasizes managing infrastructure and applications through version-controlled declarative configurations, which often pairs well with self-hosted or hybrid CI/CD setups. The report also notes that CI/CD adoption is nearly universal (83.4%), indicating that the focus has shifted from *whether* to adopt CI/CD to *how* to optimize and control it. In practice, practitioners should carefully evaluate the total cost of ownership and operational overhead of self-hosting versus managed services, considering their specific security, compliance, and customization needs. For those already self-hosting, this trend validates their approach and may lead to more robust tooling and community support for self-managed solutions. Regarding AI, teams should explore how AI can augment existing CI/CD processes, particularly in areas like intelligent log analysis, predictive failure detection, and automated test case generation, even if direct in-pipeline execution remains nascent. The report suggests that AI is primarily used *around* the pipeline for tasks like code review and test fixing, indicating that these are immediate areas for practical AI integration. Organizations should also monitor the evolving landscape of AI agents and their integration with CI/CD tools, as the report notes a significant portion of organizations are already interacting with CI/CD via agent skills.