CIQ, the enterprise software company behind Rocky Linux and the Fuzzball AI and HPC orchestration platform, announced full multi-cloud support for Fuzzball across CoreWeave, Amazon Web Services (AWS), Google Cloud Platform (GCP), Oracle Cloud Infrastructure (OCI) and Microsoft Azure. Enterprise teams define an AI training, inference or HPC workflow once and execute it across any of these environments or on-premises infrastructure, with Fuzzball routing each job automatically to the optimal destination based on cost, performance and data locality.
Enterprise AI and HPC teams pay a compounding price for every cloud (or system) they run on: rebuilt pipelines, rewritten deployment scripts, profiling, testing and validation, before a single workload can run on a new infrastructure. That cost scales directly against the speed the business demands. Fuzzball eliminates it and completely levels the playing field.
A genomics team that validates a sequencing pipeline on AWS moves it to Azure or OCI without modifying a single line in the workflow definition. A model training job that requires H100 density routes to CoreWeave automatically, while a data-sensitive simulation stays on-premises by policy. The workflow definition, container images, data orchestration and job sequencing remain identical across every environment. Teams reach production faster, access better GPU capacity at lower cost and carry no operational overhead for every cloud they add.
Also Read: Beacon.li Launches Implementation Studio, the First AI Platform to Execute Enterprise Software Implementations End to End
“AI teams today are asked to ship faster, control costs and maintain sovereignty over their data, simultaneously, across infrastructure that was never designed to work together,” said Gregory Kurtzer, CEO and founder of CIQ. “We built Fuzzball to solve that problem at the architectural level. When your workflow definition abstracts its requirements properly, you get portable access to every GPU environment the market offers and the freedom to route to wherever the best price, performance and data policy lives. Controlling your infrastructure and workloads is what enterprise AI infrastructure requires for production, and no other platform delivers it.”
One control plane across five clouds and on-premises
Fuzzball’s multi-cloud architecture rests on a provider-agnostic workflow definition. The file that describes compute jobs, data movement, container images and resource requirements carries no cloud-specific logic. Fuzzball’s orchestration layer translates that definition into concrete infrastructure on whichever environment sits underneath, whether that means Google Cloud, Microsoft Azure, Oracle Cloud, AWS or CoreWeave.
Fuzzball federates across all five cloud environments, alongside on-premises clusters. It simultaneously evaluates available environments at runtime and routes each job to its optimal destination. Enterprises gain the GPU density of CoreWeave, the breadth of three major hyperscalers and the sovereignty of on-premises infrastructure from one control plane, with no separate toolchains, deployment scripts or IAM models per provider.
One security model across every environment
Each cloud deployment is provisioned through a two-phase automated process that stands up a complete, production-ready cluster without manual intervention. Fuzzball maintains one IAM model, one set of RBAC policies and one secrets management posture across every cloud it runs on. Static credentials are eliminated at every layer: Workload Identity on GCP, Managed Identities on Azure, Dynamic Groups on OCI and IAM Roles on AWS. Security and compliance posture travel with the workflow, not the cloud.
“Fuzzball turns multi-cloud from a liability into a competitive advantage,” said Bjorn Hovland, president of CIQ. “Five clouds used to mean five IAM models, five deployment pipelines and five sets of operational overhead, with complexity and risk being multiplied.”
Source: PRNewswire