An Intellyx KubeCon 2026 Update by Jason English
It was good to catch up with Zesty at KubeCon NA 2026, because a lot has changed since our last visit just last year in 2025, as the AI revolution has changed the way clusters need to be provisioned for AI workloads.
Zesty’s latest agent sits at the application layer like a warehouse floor manager, analyzing workloads and using bin packing to assign and schedule multiple fractional workloads atop cloud resources, and now, GPUs.

Kubernetes and related projects like Karpenter don’t naturally know how to address or find fractional GPU capacity by themselves. NVIDIA has contributed several new open source plugins like RunAI that create an opportunity for clever developers to address GPUs in optimized ways like we already did with ephemeral microservices and autoscaling. Now clever vendors can crack the code of GPU optimization for AI cost and performance metrics.
“We are looking at the real usage of applications, and the desired system or GPU accommodation for each workload to try to autoscale and prevent fluctuations, but we also make sure the allocation won’t change if there is any risk of it affecting an SLA for the application,” said Shay Cohen, Group Product Manager, Zesty.
Copyright ©2026 Intellyx B.V. Intellyx is an industry analysis and advisory firm focused on enterprise digital transformation. Covering every angle of enterprise IT from mainframes to artificial intelligence, our broad focus across technologies allows business executives and IT professionals to connect the dots among disruptive trends. At the time of writing, Zesty is not an Intellyx customer. No AI was used to write this article. To be considered for a Brain Candy article, email us at pr@intellyx.com.


