Coming Soon
NVIDIA HGX RUBIN GPU SERVERS
Next-generation Rubin architecture GPU systems designed for AI training and hyperscale workloads. Plan your deployment now and be ready when systems ship.
Plan Your Rubin Deployment Explore Upcoming SystemsOVERVIEW
THE NEXT GENERATION OF AI INFRASTRUCTURE
The NVIDIA HGX Rubin platform is the successor to Blackwell, built on an entirely new GPU architecture paired with the NVIDIA Vera CPU. It is designed for the scale and complexity of next-generation AI workloads, with expected improvements in compute throughput, memory bandwidth, and system-level efficiency.
Arc Compute is working with OEM partners to prepare HGX Rubin systems for deployment. While systems are not yet available for order, our team is actively supporting early infrastructure planning, architecture design, and capacity forecasting for organizations that want to be ready when Rubin ships.
UPCOMING SYSTEMS
NVIDIA HGX RUBIN NVL8 SYSTEMS
FROM SUPERMICRO AND AIVRES
Arc Compute is preparing to offer HGX Rubin NVL8 platforms from both OEM partners. System configurations, timelines, and availability will be announced as hardware becomes commercially available.
Supermicro HGX Rubin NVL8
Coming Soon
- 2U liquid-cooled server with 8 NVIDIA Rubin GPUs and NVLink 6
- Supports NVIDIA Vera CPUs or next-generation x86 processors
- Supermicro DCBBS with Liquid-to-Air sidecar CDU option
- Designed for dense AI rack deployments scaling to 72 GPUs per rack
Aivres KR2288 HGX Rubin NVL8
Coming Soon
- 2U liquid-cooled server with 8 NVIDIA Rubin GPUs and NVLink 6
- AI-native system design optimized for high-density GPU deployments
- Direct liquid cooling for sustained high-utilization operation
- Built for large-scale AI training and production inference
Looking for multiple
systems or a full cluster?
Plan your rubin deployment
PLATFORM SPECS
NVIDIA HGX RUBIN NVL8 PLATFORM
The HGX Rubin NVL8 integrates 8 Rubin GPUs with up to 288 GB HBM4 per GPU, connected by NVLink 6 at 28.8 TB/s total bandwidth. Paired with NVIDIA Vera CPUs or x86 processors, each 2U server delivers up to 400 PFLOPS of NVFP4 inference performance. The platform can be paired with NVIDIA Vera CPUs or x86-based CPU baseboards, giving organizations flexibility in their deployment architecture.
RUBIN GPU ARCHITECTURE
Next-generation NVIDIA GPU architecture with up to 288 GB HBM4 per GPU, up to 22 TB/s memory bandwidth, and hardware-accelerated Transformer Engine with NVFP4 support. Up to 50 PFLOPS per GPU.
NVLINK 6 INTERCONNECT
Sixth-generation NVLink delivering 28.8 TB/s total GPU-to-GPU bandwidth across all 8 GPUs. Enables the entire server to operate as a tightly coupled compute domain.
VERA CPU OR X86 INTEGRATION
Flexible CPU pairing: NVIDIA Vera CPUs (88 Arm Olympus cores, 1.5 TB LPDDR5x) or next-generation x86 processors. Choose based on workload and ecosystem requirements.
LIQUID-COOLED 2U DESIGN
Full liquid cooling required for the thermal envelope of 8 Rubin GPUs at approximately 24 kW per system. Designed for dense rack deployments.
WHY ARC COMPUTE
THE TEAM BEHIND YOUR INFRASTRUCTURE
PURPOSE-BUILT INFRASTRUCTURE
Every deployment is designed around your specific workload, power, cooling, and facility requirements. No generic rack-and-ship configurations.
Trusted Across Industries
GET STARTED
PLAN YOUR NVIDIA HGX RUBIN DEPLOYMENT
Tell us about your infrastructure goals and timeline. Our team can help you evaluate Rubin-based infrastructure, plan future capacity, and prepare your organization for next-generation AI workloads.