Vultr Cloud GPUs Powered by NVIDIA A40, for Rendering and Virtual Workstations
Sep 06, 2022

The launch of our breakthrough cloud GPU platform, Talon, powered by NVIDIA was the culmination of significant research and development, and we were happy to see its positive reception covered in publications such as TechCrunch, The Next Platform, and VentureBeat. Customers have recognized the value of Vultr Talon’s "slicing" of the NVIDIA A100 Tensor Core GPU, and that our pioneering use of virtualization has enabled us to help democratize access to accelerated computing by making it more affordable than ever.

Today marks the next exciting step forward, as Vultr is the first cloud provider to introduce support for fractions of the NVIDIA A40, the world’s most powerful data center GPU for visual computing.

If you work in a studio and need render nodes for visual effects (VFX), or if you’re a creative professional who uses GPU-intensive workstation applications like computer-aided design (CAD), then Vultr Cloud GPUs powered by the NVIDIA A40 are perfect for you.

Cloud GPUs can replace your render farm or physical workstation
Artists, architects, and engineers often require substantial GPU resources to turn their visions into reality. But using an on-premises physical server or workstation is often impractical and cost inefficient, especially in the new world of normalized remote work.

The solution is to use Vultr Cloud GPUs powered by NVIDIA, which can be accessed anytime, anywhere – and deactivated when not needed, avoiding unnecessary charges.

Vultr’s virtual machines offer the ultimate in ease of use, and the NVIDIA A40 GPU is ideal for graphics workloads. Available now via our easy-to-use control panel and developer-friendly API and CLI, virtual machines with the NVIDIA A40 GPU can be provisioned in minutes.

For studios, Vultr provides on-demand cloud GPU capacity, allowing you to cut render times by accelerating your graphics pipeline. Rendering and recording software such as Blender, Maya, OBS, and Cinema 4D runs smoothly and quickly, allowing you to create 3D visualizations, animated films, and visual effects.

For creative professionals and engineers, Vultr supports top-flight virtual workstations. With support for Windows and Linux operating systems, and powered by the NVIDIA RTX Virtual Workstation (vWS) software, Vultr Cloud GPUs powered by NVIDIA deliver a low-latency user experience that is nearly indistinguishable from that of a physical workstation.

You can connect to your Windows virtual workstations using Microsoft Remote Desktop Protocol (RDP) or Parsec, the remote desktop application renowned for its high-quality game streaming. For Linux, you’ll probably want to use Parsec or VNC; here you’ll find docs relevant to Ubuntu, Debian, and Rocky Linux.

Fractions of the NVIDIA A40 GPU, starting at just $210 per month or $0.313 per hour
The NVIDIA A40 GPU is known as the most powerful data center GPU for visual computing – and for good reason. Powered by the NVIDIA Ampere architecture, one NVIDIA A40 GPU boasts rich specs:

• 48 GB GPU RAM
• 10,752 CUDA cores
• 84 RT Cores
• 336 Tensor Cores

While larger customer use cases may require one or more NVIDIA A40 GPUs, many rendering jobs or desktop apps can run quickly and efficiently with just a portion of an NVIDIA A40 GPU’s capacity.

This is why we’re introducing a range of plans using NVIDIA GPU virtualization to offer fractions of the A40 GPU. Prices start at just $210 per month, or $0.313 per hour, for an instance with 8 GB of GPU RAM.



Vultr’s fractional NVIDIA A40 GPU plans can easily save you 50% or more compared to the cost of an entire physical GPU from other cloud providers.

Bare Metal GPU for large workloads
If you need to run a workload that requires multiple physical GPUs, we also offer bare metal servers with four or eight NVIDIA A40 GPUs and dual 24-core Intel Xeon CPUs.

NVIDIA A40 and NVIDIA A100 GPUs are available now
Vultr Cloud GPUs powered by the NVIDIA A40 are available today in our New Jersey, Los Angeles, London, Sydney, and Singapore cloud data center locations, with additional capacity coming over the next few weeks in Mumbai, Frankfurt, and London. This is in addition to our offering of the NVIDIA A100 GPU, which is currently available in New Jersey, with additional capacity coming over the next few weeks in the United Kingdom, Japan, India, and Silicon Valley.

Neureveal is a Chicago-based AI company building radiology automation powered by deep learning, and it was an early adopter of Vultr Talon with the NVIDIA A100 GPU.

"Vultr’s brilliant idea to collaborate with NVIDIA to deliver the A100 GPU at a fantastic entry price point with one click scalability is a breath of fresh air,” said Dr. Abdel Fahmy, founder and CEO of Neureveal.

You can get started by provisioning Vultr Cloud GPUs powered by NVIDIA via our control panel. Additional guidance on how to get started is available in the documentation section of our website. Business users who would like a consultation with a Vultr team member or who are interested in larger-scale deployments are invited to contact us here.