ArcGIS Pro on Microsoft Azure: Deployment and Performance Guidance

Organizations across industries continue to modernize how they deliver ArcGIS Pro. As GIS workflows become more complex and organizations increasingly adopt cloud and hybrid infrastructure, Microsoft Azure provides several options for balancing performance, scalability, management, and operational efficiency.

ArcGIS Pro can be delivered through dedicated GPU-enabled virtual machines, Azure Virtual Desktop (AVD), Windows 365 GPU Cloud PCs, or Azure Local. Selecting the right approach depends not only on the number of users, but also on the GIS workflows being performed and where the supporting data and infrastructure reside.

Choosing the Right Azure Platform

Selecting the right Azure platform for ArcGIS Pro starts with four considerations:

1. Workload requirements: 2D mapping, editing, 3D visualization, imagery, geoprocessing, and GPU-accelerated analysis can place very different demands on infrastructure.
2. Data governance and location: Determine where geodatabases, imagery, LiDAR, and other GIS data must reside and how users will access it.
3. User concurrency: Consider whether users require dedicated resources or whether infrastructure can be shared across a multi-session environment.
4. Operational ownership: Determine whether your organization wants direct control over Azure infrastructure or prefers a more managed desktop experience.

These considerations help determine whether dedicated Azure GPU VMs, Azure Virtual Desktop, Windows 365 GPU Cloud PC, or Azure Local is the best fit.

GPU-Enabled Azure VMs: Flexibility and Control

GPU-enabled Azure VMs provide the resources needed to support ArcGIS Pro visualization, analysis, and GPU-accelerated workflows while providing organizations with direct control over the virtual machine environment.

They are particularly useful for users who require dedicated resources, administrative access, custom configurations, development or testing environments, or workloads that are not well suited to shared desktop infrastructure.

For most ArcGIS Pro deployments, NVadsA10_v5 is the preferred GPU-backed VM family. NCasT4_v3 remains a strong and widely used alternative where it is already deployed or where availability, region, or other Azure considerations make it a practical choice.

Single-Session Sizing (One Analyst per VM)

NVadsA10_v5 is the preferred contemporary platform. NCasT4_v3 uses an older NVIDIA T4 GPU but remains a capable and widely deployed alternative.

These configurations are starting points and should be validated against representative ArcGIS Pro workloads.

Additional GPU Options

NC_RTXPRO6000BSE_v6 – A high-end option for demanding ArcGIS Pro workloads requiring greater GPU performance, GPU memory, and CUDA compute capability. Its newer NVIDIA RTX PRO Blackwell architecture makes it well suited for complex 3D visualization, GPU-intensive analysis, and emerging workflows that place significantly greater demands on the GPU.

NVads V710 v5 – An AMD GPU-based option for graphics and visualization workloads. ArcGIS Pro uses NVIDIA CUDA libraries for various geoprocessing and GPU-accelerated capabilities. Because CUDA is not available on AMD GPUs, this VM family may not support all ArcGIS Pro workflows. Review workflow requirements and test before deployment.

Azure Virtual Desktop: Scalable Delivery for ArcGIS Pro

Azure Virtual Desktop provides a flexible platform for delivering ArcGIS Pro to individual users or larger groups while centralizing applications, data access, and infrastructure management in Azure.

ArcGIS Pro can be deployed using either single-session (personal) or multi-session (pooled) AVD host pools. Single-session environments provide dedicated VM resources for each user and are well suited for workloads requiring predictable performance. Multi-session environments allow multiple users to share a GPU-backed VM, improving resource utilization when workloads and concurrency are appropriately sized.

Fractional GPU configurations can provide an effective balance between graphics performance and cost, particularly when VM resources are matched appropriately to the expected workload.

Multi-Session Sizing (Multiple Users per VM)

User density depends on workload, available CPU, system memory, GPU memory, storage and data performance, and concurrency. These configurations should be treated as starting points and validated with representative ArcGIS Pro workflows before production deployment.

Deployment and Storage Guidance

When deploying ArcGIS Pro in Azure, consider the complete environment—including GPU, CPU, memory, storage, network latency, and data location.

1. Select the GPU-backed VM: Start with an appropriate NVadsA10_v5 or NCasT4_v3 SKU based on the expected ArcGIS Pro workload and number of users.
2. Use high-performance storage: For the best performance with shared file-based GIS data, Azure NetApp Files (ANF) is recommended. ANF provides the low-latency, high-performance shared storage needed for demanding ArcGIS Pro workflows.
3. Keep compute and data close: Deploy ArcGIS Pro VMs and ANF in the same Azure region and minimize network latency between compute and storage. Data locality is an important component of overall ArcGIS Pro performance.
4. Size storage for the workload: Configure ANF capacity and performance based on the expected workload, including latency, throughput, IOPS, concurrent users, and the characteristics of the GIS data.
5. Validate before scaling: Test representative ArcGIS Pro projects and workflows from the virtual desktop environment before production deployment. Use the ArcGIS Pro Performance Assessment Tool (PAT) where appropriate to establish a performance baseline.

Why GPU Acceleration Matters

GPU acceleration is an important component of a well-performing virtualized ArcGIS Pro environment. ArcGIS Pro uses the GPU extensively for rendering and visualization, particularly with 3D scenes, imagery, complex symbology, and other graphics-intensive workflows.

GPU capabilities are also increasingly important beyond visualization. Some ArcGIS Pro geoprocessing and analysis workflows use NVIDIA CUDA libraries for GPU acceleration, making GPU architecture, available GPU memory, and driver compatibility important considerations when selecting a virtual machine.

For more demanding workloads, additional GPU resources may provide significant benefits. This is particularly important for complex 3D visualization, imagery, GPU-compute workflows, and emerging capabilities that place greater demands on GPU compute and memory.

Windows 365 GPU Cloud PC: Turnkey Simplicity

For users who need ArcGIS Pro without managing the underlying virtual desktop infrastructure, Windows 365 GPU Cloud PC provides a secure, Microsoft-managed desktop with GPU acceleration.

Windows 365 GPU Cloud PCs are workload-focused configurations, with Microsoft managing the underlying infrastructure and GPU hardware. For organizations that require specific Azure VM or GPU configurations, Azure Virtual Desktop provides greater control.

Windows 365 GPU Cloud PCs provide a simplified option when ease of deployment and management are priorities. For workloads requiring specific GPU hardware, greater infrastructure control, or detailed performance tuning, consider Azure Virtual Desktop or dedicated Azure GPU VMs.

Azure Local: Bringing Azure Capabilities On-Premises

For organizations that need to keep data and compute resources on-premises, Azure Local provides an Azure-managed infrastructure platform that can support GPU-accelerated ArcGIS Pro virtual desktops close to GIS data.

Azure Local supports both GPU Partitioning (GPU-P) and Discrete Device Assignment (DDA) for supported NVIDIA GPUs. For ArcGIS Pro virtual desktop environments, GPU-P provides an efficient way to share supported GPU resources across multiple virtual machines, while DDA can provide dedicated access to a physical GPU when required.

Running ArcGIS Pro close to file geodatabases, imagery, LiDAR, and other large datasets can reduce dependency on WAN connectivity and improve data access performance. This makes Azure Local particularly useful for organizations with data residency requirements, limited cloud connectivity, or existing on-premises infrastructure that want to retain Azure-based management and deployment capabilities.

Performance and User Experience

ArcGIS Pro performance in a virtual environment depends on the complete system. CPU performance, GPU resources, memory, storage, network latency, data location, and workload characteristics all contribute to the user experience.

To maintain a responsive ArcGIS Pro environment:

• CPU: Favor modern processors with strong per-core performance and approximately 3.0 GHz or greater clock speeds where possible.
• GPU: Size GPU and VRAM resources for the expected workload. More demanding 3D, imagery, and GPU-accelerated workflows may require additional GPU memory and compute capability.
• Memory: Provide sufficient system memory for ArcGIS Pro, the operating system, supporting applications, and expected workloads.
• Storage: Use high-performance storage and keep storage and compute resources close together. For shared file-based GIS data, Azure NetApp Files is recommended for best performance.
• Network: Minimize latency between users, virtual desktops, and GIS data. Interactive ArcGIS Pro workflows can be sensitive to network latency, particularly when accessing file-based data.
• Testing: Use representative ArcGIS Pro projects and workflows to validate the environment before scaling. The ArcGIS Pro Performance Assessment Tool (PAT) can help establish repeatable performance baselines and compare VM configurations.

No single VM specification guarantees ArcGIS Pro performance. Evaluate the complete environment and test representative workloads before production deployment.

Bringing It All Together

Whether supporting a single GIS professional or a larger virtual desktop environment, Microsoft Azure provides multiple options for delivering ArcGIS Pro with the GPU resources, scalability, and management capabilities required for modern GIS workflows.

Dedicated GPU-enabled Azure VMs provide maximum infrastructure control, Azure Virtual Desktop provides flexible single- and multi-session delivery, Windows 365 GPU Cloud PC simplifies desktop management, and Azure Local extends Azure-managed infrastructure to on-premises environments.

Successful deployments should consider the complete environment—including CPU, GPU, memory, storage, network, data location, and workload characteristics—and validate representative ArcGIS Pro workflows before scaling.