Rackrr Global Infrastructure
Rackrr provides access to a wide range of NVIDIA GPUs through our global network of providers and data centers. From flagship data center GPUs for large-scale AI training to cost-effective consumer GPUs for inference and experimentation, our marketplace offers flexible, on-demand GPU compute.
Pricing and availability are dynamic — visit the Rackrr platform for real-time rates and to provision instances instantly.
Data Center / HPC GPUs
High-performance GPUs designed for large-scale AI training, inference, and HPC workloads.
| GPU Model | Architecture | VRAM | Memory Type | Memory BW | Details |
|---|---|---|---|---|---|
| NVIDIA® B200 | Blackwell | 192 GB | HBM3e | 8 TB/s | View → |
| NVIDIA® H200 | Hopper | 141 GB | HBM3e | 4.8 TB/s | View → |
| NVIDIA® H200 NVL | Hopper | 141 GB | HBM3e | 4.8 TB/s | View → |
| NVIDIA® H100 SXM | Hopper | 80 GB | HBM3 | 3.35 TB/s | View → |
| NVIDIA® H100 NVL | Hopper | 94 GB | HBM3 | 3.9 TB/s | View → |
| NVIDIA® A100 SXM4 | Ampere | 80 GB | HBM2e | 2.0 TB/s | View → |
| NVIDIA® A40 | Ampere | 48 GB | GDDR6 ECC | 696 GB/s | View → |
Professional / Workstation GPUs
Enterprise-grade GPUs built for AI development, rendering, and professional visualization.
| GPU Model | Architecture | VRAM | Memory Type | Details |
|---|---|---|---|---|
| NVIDIA® RTX PRO 6000 WS | Blackwell | 96 GB | GDDR7 | View → |
| NVIDIA® RTX PRO 6000 Server | Blackwell | 96 GB | GDDR7 | View → |
| NVIDIA® RTX PRO 4000 | Blackwell | 20 GB | GDDR7 | View → |
| NVIDIA® RTX A5000 | Ampere | 24 GB | GDDR6 | View → |
| NVIDIA® RTX A4000 | Ampere | 16 GB | GDDR6 ECC | View → |
Consumer / Enthusiast GPUs
High-performance consumer GPUs offering excellent price-to-performance for AI inference, training, and experimentation.
| GPU Model | Architecture | VRAM | Memory Type | Memory BW | Details |
|---|---|---|---|---|---|
| NVIDIA® GeForce RTX 5090 | Blackwell | 32 GB | GDDR7 | 1,792 GB/s | View → |
| NVIDIA® GeForce RTX 5060 Ti | Blackwell | 16 GB | GDDR7 | 448 GB/s | View → |
| NVIDIA® GeForce RTX 4090 | Ada Lovelace | 24 GB | GDDR6X | 1,008 GB/s | View → |
| NVIDIA® GeForce RTX 4060 Ti | Ada Lovelace | 8/16 GB | GDDR6 | 288 GB/s | View → |
| NVIDIA® GeForce RTX 3090 | Ampere | 24 GB | GDDR6X | 936 GB/s | View → |
| NVIDIA® GeForce RTX 3080 | Ampere | 10/12 GB | GDDR6X | 760 GB/s | View → |
| NVIDIA® GeForce RTX 3060 | Ampere | 12 GB | GDDR6 | 360 GB/s | View → |
| NVIDIA® GeForce RTX 2080 Ti | Turing | 11 GB | GDDR6 | 616 GB/s | View → |
Training Bursts
Rackrr offers a variety of HPC stock on a short-term basis for use cases like training bursts. Check the available inventory here.