Compact entry-level Ada Lovelace workstation GPU for professional visualization.
The RTX 2000 Ada is available at 2 cloud providers for $0.24/hr per GPU.
Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.
Low-power 70W card with ECC memory in a compact dual-slot form factor. Fits small form-factor workstations for CAD, 3D visualization, and light inference tasks.
Limited to PCIe 4.0 x8 bandwidth and 128-bit memory bus. For heavier compute or larger models, consider the RTX 4000 Ada with 20GB and higher throughput.
The median on-demand price has held steady around $0.24/hr per GPU across providers.
With 16GB of VRAM, the RTX 2000 Ada is best for 7B-class models in 4-bit or 8-bit quantized form, and smaller models in FP16.
The RTX 2000 Ada has 16GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
The RTX 2000 Ada has 224 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
The RTX 2000 Ada supports 7 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP8, INT4, INT8.
No. The RTX 2000 Ada is a PCIe-only GPU with no NVLink, so it is better suited to single-GPU inference and smaller-scale workloads than large distributed training jobs.
RTX 2000 Ada pricing is currently $0.24/hr per GPU.
At 720 hours per month, one RTX 2000 Ada currently costs about $172.80/mo.
The RTX 2000 Ada is available from 2 cloud providers: Runpod, GPU.ai.
Yes. We currently track 3 RTX 2000 Ada listings across 2 cloud providers:
| Billing type | Listings | Avg $/GPU/hr |
|---|---|---|
| On-demand | 2 | $0.24/hr |
| Reserved | 1 | Custom |
| GPU Memory | 16GB GDDR6 |
| Memory Interface | 128-bit |
| Memory Bandwidth | 224 GB/s |
| Error Correcting Code | Yes |
| Architecture | NVIDIA Ada Lovelace |
| CUDA Cores | 2,816 |
| Tensor Cores (Fourth-generation) | 88 |
| RT Cores (Third-generation) | 22 |
| Single-Precision Performance | 12.0 TFLOPS |
| RT Core Performance | 27.7 TFLOPS |
| Tensor Performance | 191.9 TFLOPS |
| System Interface | PCIe 4.0 x 8 |
| Power Consumption | Total board power: 70 W |
| Thermal Solution | Active |
| Form Factor | 2.7" H x 6.6" L, dual slot |
| Encode/Decode Engines | 1x encode, 1x decode (+AV1 encode and decode) |
| VR Ready | Yes |
| Graphics APIs | DirectX 12, Shader Model 6.6, OpenGL 4.65, Vulkan 1.35 |
| Compute APIs | CUDA 11.6, OpenCL 3.0, DirectCompute |
| NVIDIA NVLink | No |
Higher-tier compact Ada Lovelace workstation GPU. From $0.16/hr per GPU across 4 providers.
Previous-generation entry-level workstation GPU. From $0.03/hr per GPU across 3 providers.
Ada Lovelace cloud inference GPU. From $0.11/hr per GPU across 16 providers.
Last updated
