90SKC000-MAKAN0 ASUS Nvidia RTX PRO 6000 Blackwell 96GB GDDR7 2617MHz 1600GB/s GPU

High-Performance PCIe 5.0 GPU Accelerator for AI, HPC and VDI with the 90SKC000-MAKAN0

The 90SKC000-MAKAN0 is an ASUS server-edition GPU accelerator based on the NVIDIA RTX PRO 6000 Blackwell architecture, designed for dense multi-GPU nodes where deterministic throughput and platform compatibility are essential. With 24,064 CUDA cores and Tensor Cores enabled, it targets enterprise AI inference and training pipelines while sustaining FP32 and FP64 performance for simulation and mixed-precision compute.

To reduce memory-bound bottlenecks, the 90SKC000-MAKAN0 integrates 96GB GDDR7 on a 512-bit interface, delivering 1,792 GB/s bandwidth for large models, high-resolution datasets, and VRAM-resident workloads. PCIe 5.0 x16 connectivity improves host-to-device and peer traffic in modern CPU platforms, enabling scalable expansion across multi-accelerator servers. The full-height, dual-slot design supports standard enterprise chassis layouts; a fanless cooler relies on system airflow for predictable thermal behavior. For visualization and virtualization stacks, 4x DisplayPort 2.1b UHBR20 outputs enable high-fidelity remote graphics workflows where local display connectivity is required.

Why Choose the 90SKC000-MAKAN0?

The 90SKC000-MAKAN0 delivers a balanced density-to-performance profile for data centers standardizing on PCIe 5.0, combining high VRAM capacity with Blackwell compute for consolidating AI, HPC, and graphics acceleration. Its server-focused mechanical format and power delivery simplify deployment in validated multi-GPU platforms, while enabling a clear scaling path from single-GPU inference nodes to multi-GPU training and VDI hosts.

ASUS applies an engineering approach centered on platform-level reliability, electrical margining, and integration consistency with contemporary rack server standards, supporting predictable qualification cycles and lifecycle management in enterprise environments.

Key Applications:

• AI inference and training
• Multi-GPU compute nodes
• HPC simulation and FP64 workloads
• VDI and virtual workstation graphics
• Professional visualization and rendering

At Happyware - Server Hardware you benefit from non-binding offers tailored to your deployment targets for the 90SKC000-MAKAN0, and can extend the warranty up to 6 years to align GPU lifecycle coverage with your data center service plans.