CI/CD in 2026: The Rise of Self-Hosting and AI-Driven Workflows, But Not Within the Pipeline

A recent report by JetBrains, compiling data from four studies including their 2026 CI/CD Tools Adoption Survey, reveals significant shifts in the Continuous Integration/Continuous Delivery landscape. The report highlights that CI/CD is now a near-universal practice, with 83.4% of developers running pipelines. However, a notable trend is the resurgence of self-hosted CI/CD solutions, with 72.8% of organizations running their primary CI/CD system on-premises or on self-managed cloud instances, and on-premises installations reaching a five-year high of 30.5%. This shift towards self-hosting is critical for practitioners. It signifies a move away from fully managed cloud CI services, driven by factors such as the need for greater control over infrastructure, data, security, and potentially cost optimization. For organizations, this means re-evaluating their CI/CD strategy to determine if the benefits of self-hosting outweigh the operational overhead. The report also sheds light on the role of AI in CI/CD. While nearly half (48.8%) of organizations use AI for at least one CI/CD task, only a small fraction (8%) run AI-powered steps directly inside the build or test process. This indicates that AI's impact is currently more pronounced in auxiliary functions, such as code review, test generation, and diagnosing build failures, rather than core pipeline execution. This trend aligns with the broader industry movement towards platform engineering and developer experience. As CI/CD becomes ubiquitous, organizations are seeking ways to optimize their pipelines for security, observability, and developer-friendliness. The emphasis on self-hosting reflects a desire for more granular control over the entire software delivery lifecycle, enabling teams to tailor their environments to specific needs and compliance requirements. The cautious integration of AI directly into pipelines, favoring external or supporting AI tools, suggests a pragmatic approach to adopting new technologies, prioritizing stability and control in critical build processes. In practice, this means that while AI agents are becoming more prevalent, developers are not necessarily waiting for CI/CD vendors to integrate AI directly into their products. Instead, they are connecting existing AI agents to their CI/CD tools. Practitioners should focus on leveraging AI for tasks that augment human capabilities and improve efficiency around the pipeline, such as intelligent code suggestions, automated test case generation, and predictive analytics for pipeline failures. Furthermore, teams should carefully consider the trade-offs involved in self-hosting, including the increased responsibility for infrastructure management, security, and maintenance, versus the flexibility and cost control it offers. The report also highlights the continued consolidation of toolchains, with organizations looking to reduce the number of disparate tools in their CI/CD ecosystem. This suggests a need for integrated solutions that can provide comprehensive capabilities while maintaining the flexibility for self-managed deployments and AI integration.