Research Finder
Find by Keyword
Is Kubernetes Ready for Mission-Critical Workloads?
Exploring the shift toward GPU optimization, sovereign clouds, and the struggle to simplify container sprawl for the modern enterprise.
3/27/2026
Key Highlights
The shift from basic container orchestration toward specialized GPU and AI infrastructure management.
The rising necessity of digital sovereignty and localized cloud control in European and regulated markets.
A renewed focus on virtualization migrations as organizations seek alternatives to legacy hypervisors.
The transition of cost management from simple reporting to automated, developer-centric rightsizing.
Analyst Take: KubeCon EU Trip Report
We just wrapped up our time at KubeCon EU in Amsterdam, and the atmosphere was less about "cloud native" as a buzzword and more about the gritty reality of Day 2 operations. Based on our analysis of the conversations on the ground, the industry is architected to move past the initial excitement of containers and into the complex world of high-performance AI and strict regulatory compliance. Our meetings revealed a clear trend: the market is being redesigned to handle the "AI app explosion" while trying to keep costs from spiraling out of control. The era of just "getting it to work" is over. Now, it must work efficiently, legally, and at a massive scale. If you want my on the show floor, take rather than reading on, check that out here.
The GPU and AI Infrastructure Pivot
We spent a good amount of time with Mirantis and Nutanix discussing how the stack is being optimized for the "metal to model" transition. Stephen Frassetti at Mirantis walked us through its K0rdent solution, which aims to deliver better multi tenancy for GPU workloads. This is a tough engineering hurdle because "noisy neighbor" issues at the GPU level can ruin performance for shared clusters. We also touched base with Chris Brown and Sonali Mishra at Nutanix, who are leaning into this with the company’s Agentic AI announcement. It is designed to provide tenancy management and governance specifically for these high-cost resources. The goal here is clear: make the hypervisor work harder so the AI models can run smoother. We also met with Gari Singh from Google for a quick briefing follow up on GKE updates. It is evident that the hyperscalers are feeling the pressure to make Kubernetes a first-class citizen for massive AI training runs, not just web apps.
The Sovereignty and Multi-Cloud Reality
Sovereignty was a constant theme in Amsterdam. We met with the SUSE team, including Rhys Oxenham, Peter Smails, and Abinhav Puri, to discuss how SUSE’s solution offerings around SLES and virtualization. They are seeing a major push toward anything that moves workloads toward more secure, controllable foundations. The SUSE conversation continued with Kevin Cochrane, the CMO of Vultr, as the two companies had inked a collaboration at the event. Check out our video with Kevin and our Research Note on this topic.
Maxime Hurtrel from OVHcloud reinforced that European enterprises are no longer treating sovereignty as an afterthought. It is now a primary procurement requirement. We also spoke with Dirk Alshuth, CMO at Emma, who is seeing high demand for its multi-cloud management platform because of these requirements. Paul Hinz and Tom Brightbill at VCluster discussed the value of sovereignty and how K8S cluster management is related to the overall drive for sovereignty. They understand that the cost of compliance is becoming a major line item in enterprise budgets.
Virtualization and the "VMware Flight" (or not)
The shifting landscape of the hypervisor market was a major talking point. We talked with Stu Miniman at Red Hat about the rising interest in KubeVirt and OpenShift and migrations away from what they term as legacy virtualization. Organizations are looking for a pragmatic bridge that allows them to run VMs and containers on a single platform. We also sat down with the VMware team, including Timmy Carr, Himanshu Singh, and Dilpreet Bindra, to discuss how they see that VMware is critical for K8S deployments and how they are innovating in the open to make this happen. The tension in this space is palpable. While Broadcom iterates on its strategy, the ecosystem is moving quickly to provide alternatives that feel more "open." Jesse Butler from the AWS serverless team also joined the fray to talk about managed K8S and serveless innovation and the work they are doing in this space. Our YouTube video is a must-watch.
Automation Over Observation in FinOps
Cost management is moving from reporting to remediation. We met with Omer Hamerman and the leadership at Zesty. The company’s proprietary operator, Kompass, is designed to handle pod rightsizing and node scaling automatically. The are moving beyond just being a KubeCost competitor by offering an automated commitment layer and autoscaling. On the other side, Joe Dahlquist at IBM is seeing explosive growth since the company’s acquisition by IBM. The message from the show floor was that manual rightsizing is a failing strategy. If the system does not scale itself down, the human won't do it either.
Security and the Software Supply Chain
Finally, we checked in with JFrog and Minimus on the state of the software pipeline. Tal Zafrati at JFrog (Genefa you were missed) noted that while AI helps developers write more code, it creates a massive administrative burden for security teams. The company’s goal is to provide a horizontal governance layer for this explosion of binary artifacts. John Morello at Minimus is seeing 500% growth in image pull volume, driven by the need for FedRamp and DORA compliance.
Observability vendors in full force
We also met with the Dynatrace team of Jay Livens, and Jonathan Norris, to talk about reimagining the word "observability" through Open Feature and the DevCycle acquisition the company has made. Similarly, Courtney Ganon and Greg Leffler at Splunk are focusing on OpenTelemetry and eBPF instrumentation to streamline how Kubernetes environments are managed. They are aiming to deprecate older collectors in favor of more native, high-performance data streams like OTelP. Check out my Research Note on this topic here.
Looking Ahead
Based on what we are observing, the Kubernetes market is entering a phase of deep specialization. The general-purpose platforms of the last decade are being augmented by tools designed to solve very specific problems like GPU time slicing or sovereign data residency. The key trend that we are going to be tracking is the "intelligence" of the infrastructure itself. We are moving away from static configurations toward systems that use agentic logic to manage their own health, cost, and security. It is a necessary evolution to combat the sheer volume of data being generated.
Based on HyperFRAME's analysis of the market, our perspective is that the "VM to Container" migration is no longer a slow burn; it is a priority. The market is looking for a safe harbor, and the vendors who can provide a seamless transition for legacy workloads while simultaneously supporting new AI models will win the most ground. We see a massive opportunity for platforms that can bridge the gap between the old world of fixed virtual machines and the new world of fluid, AI-driven containers. This isn’t code for moving away from Broadcom though. They have a lot to offer the cloud native crowd.
Going forward, we are going to be tracking how companies perform on the "Day 2" operations side of things. It is easy to start a cluster, but it is hard to keep it secure and cost-effective for five years. The focus on container image security and automated rightsizing shows that the industry is finally admitting that the "do it yourself" era of Kubernetes is over. HyperFRAME will be tracking how the vendors we met in Amsterdam deliver actual outcomes rather than just raw tooling in future quarters. The focus has shifted from the "how" of orchestration to the "why" of business value, sovereignty, and cost efficiency.
Steven Dickens | CEO HyperFRAME Research
Regarded as a luminary at the intersection of technology and business transformation, Steven Dickens is the CEO and Principal Analyst at HyperFRAME Research.
Ranked consistently among the Top 10 Analysts by AR Insights and a contributor to Forbes, Steven's expert perspectives are sought after by tier one media outlets such as The Wall Street Journal and CNBC, and he is a regular on TV networks including the Schwab Network and Bloomberg.