ND128isr_NDR_GB200_v6 by Microsoft Azure
(All-cores)
(Single-core)
Specifications #
Server Metadata #
Vendor ID | azure |
Server ID | Standard_ND128isr_NDR_GB200_v6 |
Name | ND128isr_NDR_GB200_v6 |
API Reference | Standard_ND128isr_NDR_GB200_v6 |
Description | NDISRGB200V6NDR family [Isolated] [Premium Storage capable] [RDMA capable] [Intel processor] 128 vCPUs |
Family | GPU Accelerated |
Series | NDISRGB200V6NDR |
Hw Virt | - |
Status | active |
Observed At | 2026-11-21T04:48:31.297140 |
Availability #
| Region / ID | Spot | Ondemand |
|---|---|---|
East US (US) / eastus | 108.16 USD/h | 108.16 USD/h |
East US 2 (US) / eastus2 | 108.16 USD/h | 108.16 USD/h |
West US 2 (US) / westus2 | 108.16 USD/h | 108.16 USD/h |
West US 3 (US) / westus3 | 108.16 USD/h | 108.16 USD/h |
Korea South (KR) / koreasouth | 124.4 USD/h | 124.4 USD/h |
Indonesia Central (ID) / indonesiacentral | 126.54 USD/h | 126.54 USD/h |
Canada East (CA) / canadaeast | 129.8 USD/h | 129.8 USD/h |
North Central US (US) / northcentralus | 129.8 USD/h | 129.8 USD/h |
Central US (US) / centralus | 133.037 USD/h | 133.037 USD/h |
UK South (GB) / uksouth | 135.2 USD/h | 135.2 USD/h |
Germany West Central (DE) / germanywestcentral | 140.6 USD/h | 140.6 USD/h |
Italy North (IT) / italynorth | 140.6 USD/h | 140.6 USD/h |
Spain Central (ES) / spaincentral | 140.6 USD/h | 140.6 USD/h |
Korea Central (KR) / koreacentral | 146 USD/h | 146 USD/h |
Australia Southeast (AU) / australiasoutheast | 151.44 USD/h | 151.44 USD/h |
Australia East (AU) / australiaeast | 156.84 USD/h | 156.84 USD/h |
South India (IN) / southindia | 163.32 USD/h | 163.32 USD/h |
UAE Central (AE) / uaecentral | 201.16 USD/h | 201.16 USD/h |
Brazil South (BR) / brazilsouth | 216.32 USD/h | 216.32 USD/h |
Processor #
vCPUs | 128 |
Hypervisor | Microsoft Hyper-V |
CPU Allocation | Dedicated |
CPU Architecture | arm64 |
System Resources and Accelerators #
| Memory | |
|---|---|
Memory Amount | 864 GiB |
| Accelerators | |
|---|---|
Accelerator Count | 4 |
Accelerator Memory Min | 192 GiB |
Accelerator Memory Total | 768 GiB |
Accelerator Type | GPU |
Accelerator Manufacturer | NVIDIA |
Accelerator Family | Blackwell |
Accelerator Model | GB200 |
Accelerators |
| Storage | |
|---|---|
Storage Size | 16459 GB |
Storage Type | nvme ssd |
Storages |
|
Compatible Storage IDs | PremiumV2_LRS, Premium_LRS, Premium_ZRS, StandardSSD_LRS, StandardSSD_ZRS, Standard_LRS, UltraSSD_LRS |
| Network | |
|---|---|
Inbound Traffic | 0 GB/month |
Outbound Traffic | 0 GB/month |
IPv4 | 0 |
CPU and System Topology #
Server Description #
Quad NVIDIA Blackwell GB200 accelerators with 768 GB VRAM, 128 dedicated arm64 vCPUs, RDMA clustering, and 16 TB NVMe storage.
Microsoft Azure ND128isr_NDR_GB200_v6 is a dedicated GPU-accelerated instance designed for massive artificial intelligence and deep learning workloads. Running on Microsoft Hyper-V, the system provides 128 dedicated vCPUs on an arm64 architecture, paired with 864 GB of system memory at 6.75 GB per core. Accelerator capability is driven by four NVIDIA Blackwell GB200 GPUs providing a combined 768 GB of VRAM, targeted at high-parameter model execution. For high-throughput local operations and checkpoint retention, the instance features 16,459 GB of bundled NVMe SSD storage. Built-in RDMA support facilitates efficient multi-node scaling across distributed compute fabrics without standard network bottlenecks. With isolated physical tenancy and high memory density across both host and accelerator domains, this server is specifically optimized for large language model training, foundation model fine-tuning, and high-throughput computational AI pipelines.
Economics #
Alternatives #
Servers of the Same Family #
| INSTANCE | vCPUs | MEMORY | Accelerators |
|---|---|---|---|
| NV4ads_V710_v5 | 4 | 16 GiB | ⅙ |
| NC4as_T4_v3 | 4 | 28 GiB | 1 |
| NV6ads_A10_v5 | 6 | 55 GiB | ⅙ |
| NV6s_v2 | 6 | 112 GiB | 1 |
| NG8ads_V620_v1 | 8 | 16 GiB | ¼ |
| NV8ads_V710_v5 | 8 | 32 GiB | ⅓ |
| NC8as_T4_v3 | 8 | 56 GiB | 1 |
Similar Servers #
| INSTANCE | VENDOR | vCPUs | MEMORY | Accelerators |
|---|