Carte graphique Nvidia A16 64 Go GDDR6

The NVIDIA A16 64 GB GDDR6 is a purpose-built GPU for high-density, graphics-rich virtual desktop infrastructure (VDI) and virtual workstations. Combining NVIDIA Ampere architecture with support for NVIDIA vPC and NVIDIA RTX Virtual Workstation (vWS) software, the A16 delivers higher user density, improved multimedia encoding, and the ability to serve mixed user types on a single quad-GPU board. It supports multiple high-resolution displays and PCI Express Gen 4 connectivity for data-intensive virtualised workloads.

Advantages

The A16 provides up to double the user density versus the previous generation, enabling support for up to 64 concurrent users per board in a dual-slot form factor. Its quad-GPU design allows provisioning of mixed user profiles—virtual PCs and entry-level virtual workstations—on a single board. Hardware video encoding throughput is more than twice that of prior generations, supporting streaming and multimedia, while support for modern codecs (H.265 encode/decode, VP9, AV1 decode) ensures high-quality video. Built on NVIDIA Ampere architecture, it includes CUDA cores, second-generation RT Cores, and third-generation Tensor Cores for flexibility in hosting virtual workstations or running compute workloads. PCI Express Gen 4 provides fast data transfer, and support for multiple high-resolution monitors (up to two 4K or a single 5K) improves productivity in virtual environments.

Suitable for

• High-density VDI deployments for remote workers using NVIDIA vPC
• Entry-level virtual workstations running CAD or similar GPU-accelerated applications with NVIDIA RTX vWS
• Mixed user environments requiring both virtual PCs and virtual workstations on the same server
• Streaming and multimedia workloads that require high encoder throughput and modern codec support
• Virtualised setups needing multiple high-resolution displays

Key specifications

• Graphics processor: A16 (NVIDIA Ampere architecture)
• Memory: 64 GB GDDR6
• Interface: PCI Express x16 Gen 4
• Form factor: dual-slot, quad-GPU board
• Video codecs: H.265 encode/decode, VP9, AV1 decode
• Maximum concurrent users: up to 64 per board (graphics-rich VDI)

How to use

Install the Nvidia A16 into a compatible server PCIe x16 Gen 4 slot following the server manufacturer’s hardware guidelines. Combine the card with NVIDIA vPC software for graphics-rich virtual PCs or with NVIDIA RTX Virtual Workstation (vWS) software for virtual workstations. Configure user profiles and resource allocation through your virtualization platform to provision mixed user types and frame buffer sizes. Ensure proper passive cooling airflow in the server chassis according to system requirements.

Recommendations

For optimal VDI density and performance, pair the A16 with NVIDIA vPC or RTX vWS software, use a virtualization platform that supports GPU partitioning, and provide adequate server cooling and PCIe Gen 4 host support. Match framebuffer sizes to user profiles (virtual PC vs virtual workstation) to balance user density and application requirements.