NVIDIA Virtual GPU (vGPU) technology unlocks AI capabilities within Virtual Desktop Infrastructure (VDI), making it more powerful and versatile than ever before. By powering AI-driven workloads across virtualized environments, vGPU boosts productivity, strengthens security, and optimizes performance. The latest software release empowers businesses and developers to push innovation further, transform workflows, and maximize the potential of accelerated computing.
This post covers NVIDIA vGPU 18.0 release highlights, including expanding supported ecosystem platforms, new AI vWS toolkits, and enhancements for GPU utilization.
NVIDIA vGPU support for Microsoft Windows Server 2025
NVIDIA vGPU 18.0 offers support for Microsoft Windows Server 2025, enabling users to leverage advanced management features such as GPU partitioning and live migration across distributed environments. With these capabilities, users can efficiently manage multiple workloads, perform hardware maintenance, and upgrade software without disrupting virtual machines.
Additionally, NVIDIA vGPU 18.0 supports Windows Subsystem for Linux (WSL) with Windows Server 2025, enabling seamless execution of Linux distributions alongside Windows. This enhancement makes it an ideal solution for software development, testing, and running Linux-based applications in nested virtualization.
Proxmox VE, the newest supported KVM platform
Starting with NVIDIA vGPU 18.0, Proxmox VE, an open-source server management platform for enterprise virtualization, is part of the NVIDIA vGPU supported hypervisors. Users can manage demanding Linux and Windows application workloads with software-defined GPU resources with the Proxmox platform.
“Deploying NVIDIA vGPU software on Proxmox VE is seamless,” said Daire Byrne, global head of systems at DNEG, an award-winning creative studio. “The integration makes it easy to verify the performance needed to support various compute-intensive workloads and manage GPU utilization efficiently. Now we can quickly adjust resources to ensure an optimal user experience.”
Check out the step-by-step installation of NVIDIA vGPU on Proxmox VE. To learn more, read the Proxmox press release.?
New AI vWS Toolkit for fine-tuning and customizing LLMs
NVIDIA introduced in November 2024 the NVIDIA AI Virtual Workstation (vWS) Toolkits, available with the NVIDIA RTX Virtual Workstation software. These toolkits enable organizations to accelerate AI development through VDI. A new toolkit is now available: Fine-Tuning and Customizing LLMs enables developers to use parameter-efficient fine-tuning techniques (PEFT) like low-rank adaptation (LoRA) to attach custom adapters to a single large language model (LLM) for various use cases.
AI-ready with NVIDIA vGPU and Citrix
NVIDIA and Citrix have partnered to deliver AI Virtual Workstations, enabling organizations across industries such as healthcare, financial services, and government to expand their investment and use cases in VDI to accelerate AI-driven business objectives. This collaborative initiative enables Citrix-powered virtualized environments to run AI workloads while centralizing data and IT management to ensure security, scalability, and performance. For more details about this partnership, read the Citrix press release.
Heterogeneous vGPU support for a wider range of GPUs
Heterogeneous vGPU enables a single GPU to support multiple vGPU profiles (A, B, Q) with varying frame buffer sizes concurrently. With NVIDIA vGPU 18.0, this feature is extended to the NVIDIA Turing and Volta GPUs that are compatible with NVIDIA vGPU. Users can optimize GPU utilization across multiple generations of accelerated computing.
Coming soon: vGPU support for NVIDIA Blackwell GPUs
At NVIDIA GTC 2025, NVIDIA unveiled the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, the ultimate universal data center GPU equipped with 96 GB of ultra-fast GDDR7 memory and the multi-instance GPU (MIG) capability. Paired with NVIDIA vGPU software, the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU can dynamically partition compute resources for every workload, from graphics to AI. The vGPU support for NVIDIA RTX PRO 6000 Blackwell Server Edition GPU is expected in the latter half of 2025.
Get started
NVIDIA vGPU 18 enables more virtualized platforms to benefit from accelerated computing, enhancing the value of your existing investment. Ready to get started? Download now or request a free 90-day trial license. Access NVIDIA AI vWS Toolkits to start developing AI applications.
Join us for NVIDIA GTC 2025 and check out the session, Heritage Meets Technology: Leveraging Virtualized Platforms for Architectural Conservation and Business Innovation.
