A new rack-scale platform gives Cirrascale customers more room to grow, without locking them into a closed stack.
Most teams building serious AI infrastructure eventually run into the same fork in the road. Commit early to one accelerator vendor and one fabric, and scaling gets simpler but narrower. Stay open, and every workload gets matched to the hardware that actually fits it, at the cost of more moving parts to manage. Cirrascale has built its cloud around solving that second problem for customers, and it just gave itself a lot more to work with.
Cirrascale Cloud Services has announced support for the AMD Helios rack-scale solution and AMD Instinct MI400 Series GPUs across its AI innovation cloud. The addition brings AMD rack-scale infrastructure, powered by AMD Instinct MI455X GPUs, to customers running large-scale inference, frontier-model training, and fine-tuning, and it extends the open, multi-vendor approach that has defined Cirrascale's platform from the start.
What AMD Helios adds to the rack
AMD Helios is a rack-scale AI infrastructure platform built for frontier AI development and AI factory deployments. It brings together AMD Instinct GPUs, 6th Gen AMD EPYC CPUs, AMD Pensando networking, and AMD ROCm software in a single 72-GPU rack, built from OCP ORW-aligned 4-GPU compute trays. Defense-in-depth security, virtualization, and rack-scale software acceleration are built into the design, so customers can deploy on open standards and scale from a single rack to full cluster without switching architectures along the way.
At rack scale, Helios is built for exaflop-class performance: up to 2.9 exaFLOPS of peak 4-bit (OCP MXFP4) compute and 31 TB of HBM4 memory in a single rack. It scales up inside the rack over UALink over Ethernet, and scales out across racks and clusters on standards-based Ethernet aligned with the Ultra Ethernet Consortium. That matters as much as the raw numbers do: customers get a path to grow their footprint without getting tied to a proprietary fabric to do it.
MI430X: sovereign AI and HPC on open infrastructure
The AMD Instinct MI400 Series also includes the AMD Instinct MI430X GPU, built specifically for sovereign AI and HPC workloads. It pairs leadership performance for simulation and modeling with AI capability in systems that are open and auditable rather than closed. For governments, research institutions, and national labs, that combination offers a way to scale private AI compute while keeping control over how and where it runs.
Why this fits how Cirrascale builds
Cirrascale delivers these platforms the way it delivers everything else on its AI innovation cloud: as dedicated, bare-metal capacity, backed by hands-on expert support. That model gives customers room to match the accelerator to the workload instead of standardizing on one vendor because switching later would be too costly. It is also a natural fit for how AMD built Helios in the first place, on open standards and standards-based Ethernet, designed for interoperability rather than lock-in.
“Our customers are building at a scale where infrastructure choices carry real weight, and they want to match the accelerator to the workload instead of being locked into one path. The AMD Helios rackscale solution fits how we think about the cloud: open, standards-based, and built to run the most demanding inference and training jobs without forcing customers into a closed system.” -- Alex Nataros, Chief Technology Officer, Cirrascale Cloud Services
“AMD Helios rackscale solution brings the best of AMD compute and networking technology together into a single rack-scale building block, delivering leadership performance, performance per watt, and cost per token for the next wave of AI. As demand accelerates for large-scale inference and frontier-model training, AMD Helios gives customers a clear path to scale from rack to cluster with the efficiency, economics, and openness needed to power the most demanding AI workloads.” -- Andrew Dieckmann, Corporate Vice President and General Manager, Data Center GPU Business Unit, AMD
What's next
Cirrascale expects to make AMD Helios and AMD Instinct MI400 Series capacity available on its AI innovation cloud as the platform reaches general availability. For a neocloud built around Private AI, it's another way to give customers more of what they've been asking for: real choice in how they build, without giving up the support and control that come with dedicated infrastructure.
Interested in early access, capacity planning, or pricing? Visit cirrascale.com/ai-innovation-cloud/amd-instinct-series-cloud to learn more.
.jpg)
