![NVIDIA RTX™]()
NVIDIA RTX™
from AMBER AI & Data Science Solutions
NVIDIA RTX PRO™ 6000 Blackwell
The most powerful professional-GPU family for workstation and server-class AI.
The NVIDIA RTX PRO™ 6000 Blackwell extends the NVIDIA professional GPU portfolio to the Blackwell generation, available in three editions, Workstation, Max-Q, and Server, for desktop, compact / mobile workstation, and data-centre deployments.
The Workstation Edition ships with 96 GB of GDDR7 memory, 5th-generation Tensor Cores delivering 4,000 TOPS of AI performance, 4th-generation ray-tracing cores at 380 TFLOPS, PCIe Gen 5, and four DisplayPort 2.1 outputs.
AMBER configures RTX PRO 6000 Blackwell systems for European enterprises and research institutes that need professional-GPU capability alongside their AI infrastructure, the same audience that deploys AMBER's NVIDIA DGX™ and AI Factory systems.
Workstation Edition specifications
| GPU Architecture | NVIDIA Blackwell |
|---|---|
| AI TOPS | 4,000 |
| Tensor Cores | 5th Generation |
| Ray Tracing Cores | 4th Generation |
| NVIDIA Encoder (NVENC) | 4× 9th Generation |
| NVIDIA Decoder (NVDEC) | 4× 6th Generation |
| Memory Configuration | 96 GB GDDR7 with error-correcting code (ECC) |
| Memory Bandwidth | 1,792 GB/sec |
| Max Power Consumption | 600 W |
NVIDIA RTX PRO™ 6000 Blackwell Max-Q Workstation Edition
Density-optimised desktop GPU. Same Blackwell performance as the Workstation Edition, half the power envelope, designed to scale 1–4 GPUs per workstation.
The Max-Q Workstation Edition brings the full Blackwell GPU class into a 300 W power envelope optimised for high-density desktop workstations. Same 96 GB of GDDR7 memory and 5th-generation Tensor Cores as the Workstation Edition; half the maximum power consumption.
A 4.4″ × 10.5″ dual-slot PCIe card, sized to fit 1–4 GPUs in a single workstation chassis, up to 384 GB combined memory across a four-GPU configuration. PCIe Gen 5 connectivity and four DisplayPort 2.1 outputs.
AMBER configures Max-Q systems for European AI and data-science teams running multi-GPU local development pipelines where rack-mount data-centre infrastructure isn't available, the same audience that deploys AMBER's NVIDIA DGX™ systems.
Max-Q Workstation Edition specifications
| GPU Architecture | NVIDIA Blackwell |
|---|---|
| AI TOPS (FP4, theoretical, with sparsity) | 3,511 |
| Tensor Cores | 5th Generation |
| Ray Tracing Cores | 4th Generation, 333 TFLOPS |
| FP32 Single-Precision | 110 TFLOPS |
| NVIDIA Encoder (NVENC) | 4× 9th Generation |
| NVIDIA Decoder (NVDEC) | 4× 6th Generation (4:2:2 H.264 / HEVC) |
| Memory Configuration | 96 GB GDDR7 with error-correcting code (ECC) |
| Memory Bandwidth | 1,792 GB/sec |
| Max Power Consumption | 300 W |
| Form Factor | 4.4″ H × 10.5″ L, dual-slot PCIe |
| Display Outputs | 4× DisplayPort 2.1 |
| PCIe Generation | Gen 5 |
| Multi-Instance GPU (MIG) | Up to 4 fully isolated instances per GPU |
| Multi-GPU scaling | 1–4 GPUs per workstation (up to 384 GB combined memory) |
NVIDIA RTX PRO™ 6000 Blackwell Server Edition
Universal data-centre GPU for enterprise AI inference, virtual workstations via vGPU, and accelerated rendering across multi-GPU rack servers.
The Server Edition is the data-centre form factor in the RTX PRO 6000 Blackwell line. Two thermal options, both with 96 GB of GDDR7 ECC memory: dual-slot air-cooled FHFL for standard 2U / 4U rack servers, or single-slot passive liquid-cooled FHXL for dense liquid-cooled deployments. Configurable to up to 600 W.
24,064 CUDA cores, 5th-generation Tensor Cores delivering 4 PFLOPS FP4 compute, and 4th-generation RT Cores. Multi-Instance GPU (MIG) splits the card into up to four fully isolated instances; NVIDIA vGPU support for virtual workstations delivered from the data centre is on the roadmap.
AMBER configures Server Edition deployments for European enterprises building inference at scale, vGPU-delivered virtual workstations, and multi-tenant AI workloads as part of the NVIDIA AI Factory compute stack.
Server Edition specifications
| GPU Architecture | NVIDIA Blackwell |
|---|---|
| CUDA Cores | 24,064 |
| Tensor Cores | 5th Generation |
| RT Cores | 4th Generation, 188 cores, 355 TFLOPS peak |
| FP4 Tensor Core | 4 PFLOPS |
| FP8 Tensor Core | 2 PFLOPS |
| FP16 / BF16 Tensor Core | 1 PFLOP |
| TF32 Tensor Core | 234 TFLOPS |
| FP32 Single-Precision | 120 TFLOPS |
| Memory Configuration | 96 GB GDDR7 with error-correcting code (ECC) |
| Memory Interface | 512-bit |
| Memory Bandwidth | 1,597 GB/sec |
| Max Power Consumption | Up to 600 W (configurable) |
| Form Factor (Air-cooled) | Dual-slot FHFL (Full-Height, Full-Length) |
| Form Factor (Liquid-cooled) | Single-slot FHXL passive |
| Multi-Instance GPU (MIG) | Up to 4 fully isolated instances per GPU |
| Rack-server scaling | 2 / 4 / 8 GPUs per server (industry-standard 2U / 4U rack chassis) |
RTX PRO 6000 Blackwell, side-by-side
The three editions share the same Blackwell GPU class and 96 GB GDDR7 ECC memory. Form factor, power envelope, and target deployment differ.
| Specification | Workstation Edition | Max-Q Workstation Edition | Server Edition |
|---|---|---|---|
| Primary use case | Desktop workstation: design, 3D, simulation, local AI | High-density workstation: 4-GPU AI & data-science rigs | Data centre: AI inference, vGPU virtual workstations, HPC |
| Form factor | Desktop PCIe, full-size dual-slot | Desktop PCIe, 4.4″ × 10.5″, dual-slot | Air: dual-slot FHFL · Liquid: single-slot FHXL passive |
| Max power (TGP) | 600 W | 300 W | Up to 600 W (configurable) |
| Memory | 96 GB GDDR7 ECC | 96 GB GDDR7 ECC | 96 GB GDDR7 ECC |
| Memory bandwidth | 1,792 GB/s | 1,792 GB/s | 1,597 GB/s |
| CUDA cores | — | — | 24,064 |
| AI TOPS (FP4) | 4,000 | 3,511 (theoretical, sparsity) | 4 PFLOPS |
| FP32 throughput | — | 110 TFLOPS | 120 TFLOPS |
| RT Cores | 4th Gen, 380 TFLOPS | 4th Gen, 333 TFLOPS | 4th Gen, 188 cores, 355 TFLOPS peak |
| Multi-Instance GPU (MIG) | — | Up to 4 instances | Up to 4 instances |
| Display outputs | 4× DisplayPort 2.1 | 4× DisplayPort 2.1 | 4× DisplayPort 2.1 |
| Cooling | Active (fan) | Active (fan) | Air-cooled or passive liquid-cooled |
| Multi-GPU scaling | — | 1–4 GPUs per workstation | 2 / 4 / 8 GPUs per rack server |
| PCIe | Gen 5 | Gen 5 | — |
Specifications above are sourced from NVIDIA's product pages and reference documents and remain subject to NVIDIA's updates. Cells shown as — indicate values that NVIDIA does not disclose on the relevant product page or are pending AMBER verification; AMBER configures and quotes each system based on the most current NVIDIA specifications and the customer's deployment context. Talk to AMBER about the right RTX PRO 6000 Blackwell configuration →
The Future of Visual Computing for Professionals
NVIDIA RTX™ is the leading professional visual computing platform, offering a range of solutions for users to create breathtaking architectural and industrial designs, advanced special effects, and complex scientific visualizations.
With millions of creative and technical users relying on NVIDIA RTX™ to enhance their workflow, the platform transforms the challenges of today into successful solutions for tomorrow.
NVIDIA RTX™ and Quadro Solutions from Desktop to Cloud
No matter your requirements, the NVIDIA RTX™ platform provides a range of solutions to fit your workflow, including professional laptops, desktop workstations, EGX™ platforms, virtual workspaces, cloud solutions, specialty solutions, and embedded solutions.
Professional Laptops
Desktop Workstations
EGX™ Platform for Professional Visualization
Virtual Workspaces
Cloud
Embedded Solutions
Driving Innovation Across Industries
Architecture, Engineering, Construction & Operations
The NVIDIA RTX™ platform empowers millions of creative and technical professionals with the latest technology, so they can work faster, smarter, and more effectively.
Tailored solutions are available for industries such as architecture, engineering, construction and operations.
Manufacturing & Product Development
Creating innovative products of the future demands novel approaches in every aspect of development. The NVIDIA RTX™ visual computing platform, which can be accessed through physical or virtual workstations using NVIDIA vGPU software, enables cutting-edge design and engineering teams to utilize photorealistic visualization, real-time simulation, AI, and additive manufacturing to stay ahead of the competition and quickly bring innovative products to market.
Media & Entertainment
With advanced display technology, optimally tuned drivers, and exclusive features such as multi-display technology, Quadro Sync, video encoding and decoding, virtual reality, productivity tools, enterprise drivers, livestreams and video calls, NVIDIA RTX™ and NVIDIA Quadro professional GPUs elevate your workflow to a new level.
Discover the NVIDIA RTX™ possibilities!
We would be happy to advise you and create a tailor-made offer for your NVIDIA RTX™ applications.