General purpose instances featuring full passthrough on a affordable fraction of a NVIDIA A40 GPU.
The NVIDIA A40 is a high-performance data center GPU built on the NVIDIA Ampere architecture for AI training, inference, professional visualization, and virtual workstation workloads. Equipped with 48 GB of GDDR6 memory and advanced ray-tracing and Tensor Core technologies, it delivers powerful acceleration for graphics-intensive and compute-demanding applications. The A40 supports NVIDIA Virtual GPU (vGPU) software, enabling scalable virtualization and resource sharing. Its versatility, large memory capacity, and enterprise-grade reliability make it well suited for modern data center and hybrid cloud environments.
| Name | NVMe Storage | Memory | vCPUs | Transfer | Price | Regions |
|---|---|---|---|---|---|---|
G1-gpu-a40-5m-1c-2g
| 90 GiB | 5 GB | 1 vCPUs | 3 TB | $0.0959 /hr $70 /mo | 6 |
G1-gpu-a40-10m-2c-4g
| 180 GiB | 10 GB | 2 vCPUs | 4 TB | $0.1918 /hr $140 /mo | 2 |
G1-gpu-a40-20m-4c-8g
| 360 GiB | 20 GB | 4 vCPUs | 5 TB | $0.3836 /hr $280 /mo | 3 |
G1-gpu-a40-30m-6c-12g
| 550 GiB | 30 GB | 6 vCPUs | 6 TB | $0.5753 /hr $420 /mo | Sold out |
G1-gpu-a40-40m-8c-16g
| 740 GiB | 40 GB | 8 vCPUs | 8 TB | $0.7671 /hr $560 /mo | 1 |
G1-gpu-a40-60m-12c-24g
| 1110 GiB | 60 GB | 12 vCPUs | 10 TB | $1.0959 /hr $800 /mo | Sold out |
G1-gpu-a40-120m-24c-48g
| 1400 GiB | 120 GB | 24 vCPUs | 15 TB | $2.1918 /hr $1 600 /mo | Sold out |