• Posted on February 23, 2026by Angel Ramirez, CEO of Cuemby and CNCF Ambassador CNCF projects highlighted in this post Why v1.35 reads like an AI-infrastructure release Kubernetes has become the place where teams coordinate mixed production workloads: services, batch jobs, data pipelines, and ML training. • TheKubernetes v1.35 (“Timbernetes”) releasereinforces that trajectory with changes that reduce operational friction in scheduling, resource control, and configuration workflows. • What stands out in v1.35 is practical: fewer restarts for resizing, new primitives for coordinated placement, and safer defaults for how teams generate and review manifests at scale. • Taken together, these updates point to a Kubernetes control plane that’s adapting to bursty jobs, tightly coupled training runs, and continuously tuned inference services. • Teams operating mixed clusters tend to feel the pressure first in placement efficiency, resize churn, and configuration review hygiene. • The rest of this piece focuses on the v1.35 changes that ease those pressures and make AI/ML operations more predictable at scale.

Article Summaries:

  • Kubernetes v1.35, codenamed “Timbernetes,” delivers several features aimed at easing AI and ML operations. The release introduces workload‑aware scheduling, including an alpha gang‑scheduling prototype that ensures all‑or‑nothing placement for distributed training jobs. In‑place Pod resource resizing is now stable, allowing CPU and memory changes without container restarts, which reduces churn for inference services. Dynamic Resource Allocation (DRA) remains enabled to improve device‑aware scheduling. Additionally, kubectl’s default output format switches to KYAML, a stricter YAML subset designed to reduce configuration errors. These updates collectively lower operational friction for mixed production workloads that combine bursty training, steady inference, and data pipelines.
  • Kubernetes v1.35 (“Timbernetes”) reinforces the platform’s role as an AI operating system by easing operational friction for mixed workloads. Key updates include the alpha‑stage workload API and gang scheduling, enabling coordinated placement for distributed training and tightly coupled jobs. In‑place pod resource resizing moves to Stable, allowing CPU and memory adjustments without container restarts-beneficial for inference services. Dynamic Resource Allocation remains enabled to improve device‑aware scheduling. Additionally, kubectl now defaults to KYAML output, a stricter YAML subset that reduces configuration errors. Together, these changes aim to make scheduling, scaling, and configuration more predictable for AI/ML teams at scale.

Sources: