Nexus HyperFabric AI
Full-stack AI infrastructure co-developed by Cisco and NVIDIA, deployed in a few clicks.
A cloud-managed turnkey AI cluster combining Cisco networking, NVIDIA accelerated computing and AI software, and VAST storage. Build generative AI infrastructure without deep IT expertise.
What Is Nexus HyperFabric AI
Designing, configuring, and operating generative AI infrastructure from scratch requires deep expertise across networking, compute, storage, and software. Nexus HyperFabric AI absorbs that complexity — a turnkey cluster that bundles Cisco networking, NVIDIA accelerated computing and AI software, and VAST Data storage, all managed from the cloud. Guided design, deployment, and operations in the cloud console let even a small team stand up AI infrastructure.
Key Capabilities
Cisco–NVIDIA co-development
A validated full stack jointly designed by both companies (announced at Cisco Live 2024)
Integrated accelerated computing
NVIDIA Tensor Core GPUs + BlueField-3 SuperNICs/DPUs
Full stack including storage
An integrated stack including VAST Data storage, managed from the cloud
Generative AI out of the box
Inference and generative AI powered by NVIDIA NIM microservices
Components
| Layer | Composition |
|---|---|
| Network | Cisco AI Ethernet fabric (Nexus · Silicon One) |
| Compute | NVIDIA Tensor Core GPUs + BlueField-3 SuperNICs/DPUs |
| Storage | VAST Data integrated storage |
| Software | NVIDIA NIM · AI Enterprise |
| Operations | Cloud-managed (design, deployment, and monitoring unified) |
What It's Used For
Rapid AI infrastructure adoption
Stand up generative AI infrastructure quickly, even with limited specialist staff
Generative AI services
Run NIM-based internal chatbots, RAG, and inference services turnkey
Scalable clusters
Scale easily with cloud management as demand grows