In this KubeFM interview recorded live at KubeCon + CloudNativeCon North America 2025, Akamas CTO and co-founder Stefano Doni discusses the launch of Akamas Insights and why Kubernetes optimization remains one of the biggest unsolved challenges for modern teams.
Stefano explains how most organizations run heavily over-provisioned clusters while still facing performance and reliability issues – a problem rooted not only in technical complexity, but also in misaligned incentives between developers, SREs, and platform engineers. Insights was built to close this gap by turning existing observability data into safe, ready-to-apply optimization actions across the full stack: clusters, pods, and runtimes.
The conversation covers why agentless adoption matters, the limitations of tools like Cluster Autoscaler, Karpenter, VPA, and HPA, and how Akamas Insights delivers a unified, AI-powered approach that detects inefficiencies, estimates their impact, and guides teams toward the right configuration changes.
Stefano also shares Akamas’ long-term vision: enabling developers to receive optimization recommendations directly in their workflow (“shift left”), and ultimately moving toward autonomous, platform-embedded optimization (“shift down”).
Watch the interview to learn how Akamas is redefining continuous optimization for Kubernetes environments.
