The Cloud Native Computing Foundation (CNCF) announced on August 17, 2026, that the Kubeflow project had graduated, reflecting the open-source platform’s achievement of a level of technical maturity and governance that qualifies it to manage artificial intelligence and machine learning workloads in production environments on Kubernetes.
Kubeflow standardizes the stages of the artificial intelligence and machine learning lifecycle, from data processing and interactive development through distributed model training and tuning to inference and model serving. According to the CNCF, this stack can run across public, private, and hybrid cloud environments, giving organizations a unified foundation that is neutral toward any particular vendor.
Why does this matter?
The move comes as organizations shift from limited artificial intelligence experiments to operating models at scale. In practical terms, this means data science, engineering, and platform teams need a consistent architecture that moves work from experimentation to production, rather than managing separate tools for each stage of the lifecycle.
The platform targets data scientists, artificial intelligence and machine learning engineers, and platform engineering teams, in addition to organizations subject to regulatory requirements. The CNCF believes that Kubeflow’s Kubernetes-based nature helps these groups run their workloads across different environments while reducing dependence on a single platform or vendor.
Broad growth within the open-source community
Kubeflow originated at Google in 2017 and began as a collection of components before evolving into a unified platform designed around the needs of data processing and model development, training, and serving. It joined the CNCF as an incubating project in 2023.
Since then, the project has reached more than 6,600 contributors from more than 1,000 organizations, while the number of stars across its GitHub repositories has surpassed 33,000. Kubeflow packages on the Python Package Index have reached nearly 260 million downloads, with organizations such as Bloomberg, NVIDIA, Red Hat, LinkedIn, and Spotify using its subprojects to standardize artificial intelligence workloads.
Kubeflow integrates with CNCF technologies and other projects, including Prometheus for monitoring, KServe, Feast, Kueue for managing job queues, and Istio for securing communication between services.
Graduation requirements and beyond
To achieve graduated project status, Kubeflow completed a third-party security audit, established an official steering committee to support transparent governance, and adopted the CNCF Code of Conduct. The project also retains a Best Practices badge from the Core Infrastructure Initiative (CII), indicating its commitment to secure software development practices.
Chris Aniszczyk, CTO of the CNCF, said that graduation establishes Kubeflow as a mature option for enterprise artificial intelligence workloads on Kubernetes. For his part, David Aronchick, Kubeflow’s co-founder, described the milestone as the result of the project’s evolution and its adoption by teams and companies.
Next steps
The upcoming roadmap focuses on expanding orchestration for large language model (LLM) workloads and strengthening post-training capabilities, including fine-tuning, alongside large-scale data engineering and agentic workloads within the data and artificial intelligence lifecycle.
Thus, the impact of graduation is not limited to announcing a new status for the project; it gives organizations evaluating artificial intelligence architecture a signal of Kubeflow’s maturity in terms of technology, community, and governance, while keeping the platform within an open, portable framework built on Kubernetes.