The NVIDIA RTX 6000 Ada Generation is a 48GB professional workstation graphics card for rendering, AI, simulation and large visualisation workloads. Built on the NVIDIA Ada Lovelace architecture, it delivers 91.1 TFLOPS of FP32 performance in a 300W dual-slot card that fits standard workstations.
| Brand / model | NVIDIA RTX 6000 Ada Generation |
| Part number | 900-5G133-2550-000 |
| Architecture | NVIDIA Ada Lovelace |
| GPU memory | 48GB GDDR6 with ECC |
| FP32 performance | 91.1 TFLOPS (peak) |
| RT core performance | 210.6 TFLOPS |
| Tensor performance | 1457 AI TOPS (FP8, with sparsity) |
| Graphics bus | PCIe Gen 4 x16 |
| Display outputs | 4 x DisplayPort 1.4a |
| Max power consumption | 300W |
| Form factor | 4.4in (H) x 10.5in (L), dual slot |
| Thermal | Active cooling |
| Video engines | 2 x NVENC (8th gen, AV1 encode), 2 x NVDEC |
| Virtual GPU software | NVIDIA vPC/vApps, NVIDIA RTX Virtual Workstation |
| VR ready | Yes |
The card uses a PCIe Gen 4 x16 interface and a dual-slot, 10.5in long form factor with active cooling. Check that your workstation has a free x16 slot, two slots of clearance and a power supply that can deliver the card's 300W maximum. NVIDIA notes the display ports are on by default and should be turned off when using vGPU software.
Suited to: architects, engineers, 3D artists, VFX studios, data scientists and researchers who need large GPU memory for professional applications.
Not suited to: small form factor PCs, or buyers who only need light CAD or office graphics. For a newer Blackwell option, compare the NVIDIA RTX PRO 4500 Blackwell 32GB, or see all NVIDIA professional graphics cards.
48GB of GDDR6 memory with error-correcting code (ECC).
Up to 300W maximum power consumption, according to NVIDIA.
Four DisplayPort 1.4a outputs.
Yes, it supports NVIDIA RTX Virtual Workstation and NVIDIA vPC/vApps software.