Storage in the Age of AI: Rethinking Data Infrastructure

Our CEO, Rob Pankow, presented “Storage in the Age of AI: Rethinking Data Infrastructure for the Next Wave of Applications” at the Container Days London 2026 conference, held from February 11th to 12th in London, United Kingdom.

Abstract:

AI is reshaping the demands placed on data infrastructure, and Kubernetes has become the default platform where these workloads converge. Traditional storage strategies, optimized for virtual machines and monolithic applications, are breaking down under the pressure of AI workloads that demand low-latency access, massive parallelism, and seamless scalability. At the same time, enterprises face increasing pressure around cost efficiency, data governance, and developer velocity.

This talk will explore how storage architectures must evolve in the age of AI and Kuberenetes. From adopting NVMe and NVMe-oF for predictable, high-performance throughput to enabling new paradigms like database branching, ephemeral environments, and unified OLTP + OLAP workflows. We’ll discuss the pitfalls of legacy cloud storage models (IOPS throttling, over-provisioning, and cost unpredictability) and how modern approaches can eliminate these bottlenecks.

Some of the key topics covered:
– Why AI workloads on Kubernetes amplify challenges around performance, scale, and cost
– How to evaluate CSI driver capabilities for AI workloads, including Kubevirt orchestration and live migrations
– How NVMe-based architectures and
disaggregated storage unlock new efficiency gains
– What “ephemeral storage for ephemeral environments” means for testing, QA, and AI pipelines
– Practical steps to future-proof storage strategies as data becomes the backbone of intelligent applications

You can download the presentation here.