Built on Purpose. Aimed at the Future.
Gregory Kurtzer did not set out to start a company. He set out to solve a problem.
In the early 2000s, working inside Lawrence Berkeley National Laboratory, Kurtzer watched the scientists around him struggle with computing infrastructure that either didn't exist or was held together by whatever the local administrators could build themselves. So he built better tools. Warewulf, his cluster management toolkit, became foundational to some of the world's most demanding scientific computing environments. Then he co-created CentOS, the free enterprise Linux distribution that quietly became the backbone of data centers everywhere.
The pattern was the same each time: see a real need, build the right thing, give it to the people who need it.
The Problem That Founded CIQ
By the mid-2010s, containers had arrived in HPC. Scientists wanted to package genomics workflows and physics simulations into portable, reproducible environments. The problem was that the available container tools were designed for web services and enterprise applications. They clashed with the security models, performance requirements, and architectural assumptions of high-performance computing.
Kurtzer created Singularity, a container platform designed from the ground up for scientific computing. It spread quickly through the research community. But he recognized that containers were only one piece of the puzzle.
What organizations actually needed was the entire solution: a way to manage workflows, schedule jobs across distributed resources, and bridge on-premises clusters with cloud infrastructure. The HPC architecture the field relied on was thirty years old. The computing demands of modern science, and the AI workloads that were beginning to emerge, required something fundamentally different.
That realization became Fuzzball.
In April 2020, Kurtzer founded CIQ with a focused mission: build a modern, cloud-native computing platform for performance-intensive workloads. Fuzzball would be the Kubernetes of high-performance computing, an API-first, container-native architecture designed for hybrid environments, federated clusters, and the tightly coupled parallel computing that traditional schedulers were never built to support.
A Moment of Necessity
Eight months into that mission, the landscape shifted.
In December 2020, Red Hat announced that CentOS, the distribution Kurtzer had co-created, would shift from a stable downstream rebuild of Red Hat Enterprise Linux to an upstream testing ground. For thousands of organizations that depended on CentOS as their production operating system, the announcement created an immediate crisis.
Kurtzer responded the same way he always had. He posted a message to a public forum: "I'm Greg, I'm one of the original founders of CentOS, and I'm going to recreate this. Join me."
Within six weeks, over ten thousand people had. The project became Rocky Linux, named for the late Rocky McGaugh, one of Kurtzer's original CentOS co-founders. Seven months after the announcement, it shipped. Today, Rocky Linux runs on 2.75 million actively deployed instances, measured by EPEL telemetry.
Rocky Linux was not a detour from CIQ's founding vision. It was proof of something essential: that the community's need for open, stable, enterprise-grade infrastructure was exactly the problem CIQ existed to solve.
When the Wave Arrived
AI workloads, it turned out, look a lot like HPC workloads: massive parallelism, GPU-intensive computation, tight coupling between processing nodes, and enormous data movement challenges. The infrastructure CIQ had spent years designing was precisely what organizations needed as they stood up AI training and inference pipelines.
In 2023, when Red Hat further restricted access to RHEL source code, Kurtzer co-founded the Open Enterprise Linux Association alongside Oracle and SUSE, ensuring that enterprise Linux source code remained publicly available. By 2025, CIQ had secured a licensing deal making Rocky Linux the first enterprise Linux distribution authorized to include the complete NVIDIA AI software and networking stack.
From that foundation, a full product portfolio took shape. RLC Pro delivers enterprise Linux with long-term support, FIPS compliance, and direct bug fixes. RLC Pro AI is optimized for AI workloads. RLC Pro Hardened provides FIPS 140-3 compliance with hardened security configurations. Warewulf Pro handles bare-metal cluster provisioning. Fuzzball orchestrates it all.
From the kernel to the orchestrator, CIQ delivers and supports the entire stack for performance-intensive computing, including AI and machine learning workloads.
"We're finally, after almost six years, seeing it all materialize," Kurtzer says. "The same narrative, the same vision. It just took longer than anyone would have liked."
What We're Building
CIQ is the founding commercial sponsor of Rocky Linux and the company behind the infrastructure stack that research institutions, national laboratories, federal agencies, and AI-native organizations depend on to run their most demanding workloads.
The thread that runs through everything, from a national laboratory workbench to a container runtime to an operating system to a full AI infrastructure platform, is the same: build what people genuinely need, build it well, and make sure it lasts.