PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC

2,999 €

PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC

PNY

PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC graphics card

The PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC brings 5th Generation Tensor Cores, 4th Generation RT Cores with RTX Mega Geometry, and 24GB of GDDR7 ECC memory with 432GB/s of bandwidth to a dual-slot, low-profile design just 167mm long—with a TBP of 70W and no need for an external power connector. With 770 AI TOPS and 73 TFLOPS of RT Core performance, it's the most powerful low-profile professional GPU of the Blackwell generation, ready for AI inference, generative AI, and photorealistic rendering in SFF workstations, edge systems, and industrial applications.

Nvidia GeForce RTX 40

GDDR7 / 24GB

Operating memory
NVIDIA GPU

PCIe 5.0 x8

Bus interface
NVIDIA GPU

None

Additional power
Nvidia GeForce RTX 40

Blackwell (GB203)

Chip architecture
Nvidia GeForce RTX 40

550W

Recommended PSU power
PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC

Key features of the PNY NVIDIA RTX PRO 4000 SFF Blackwell 24GB GDDR7 ECC graphics card

  • The card features 24GB of GDDR7 ECC memory with 432GB/s bandwidth on a 192-bit bus for smooth work with large AI models, complex CAD assemblies, and multi-application workflows in a compact professional environment
  • NVIDIA Blackwell architecture with 8960 CUDA cores and 5th generation Tensor Cores achieves 770 AI TOPS and 24 TFLOPS in FP32—a dramatic increase over the previous Ada Lovelace generation at an identical 70W TBP
  • 4th generation RT Cores with 73 TFLOPS of performance and RTX Mega Geometry technology can process up to 100× more ray-traced triangles than the previous generation for real-time photorealistic visualisations and engineering simulations
  • 5th generation Tensor Cores with FP4 support and DLSS 4 Multi Frame Generation accelerate local LLM inference, generative AI, and AI-assisted creative tools directly on a compact workstation without cloud dependency
  • New Streaming Multiprocessors with neural shaders integrate neural networks directly into programmable shaders for next-generation AI-augmented graphics workflows, even in a half-height form factor
  • The card requires no external power connector—it draws 70W directly from the PCIe 5.0 slot, simplifying installation, reducing PSU and cable management demands, and enabling deployment in the most compact systems
  • Four mini DisplayPort 2.1b outputs let you connect four monitors simultaneously at up to 4K resolution at 165Hz, or two at 8K at 100Hz with HDR and extended colour depth for precise visualisation workflows
  • Two 9th generation NVENC video encoders and two 6th generation NVDEC decoders with support for 4:2:2 10-bit H.264 and HEVC ensure professional real-time video processing, even in a compact edge deployment

770 AI TOPS for local agentic AI with data protection

Seven hundred and seventy AI TOPS from a card that fits into any SFF PC slot—that's exactly what the NVIDIA RTX PRO 4000 24GB SFF Blackwell Low Profile delivers. 5th generation Tensor Cores with FP4 support deliver more than double the performance of the previous Ada Lovelace generation at the same 70W TBP and without the need for external power. For edge AI deployments, industrial automation systems, or design studios with compact workstations, this means local AI inference, generative AI, and LLM workflows without relying on cloud infrastructure and without needing to replace the entire PC.

NVIDIA GPU

24GB GDDR7 and RTX Mega Geometry in the smallest Blackwell design

GDDR7 memory comes to the low-profile segment for the first time with the Blackwell generation—and the NVIDIA RTX PRO 4000 24GB SFF Blackwell Low Profile takes full advantage. A total of 24GB of ECC memory with 432GB/s bandwidth can handle large CAD assemblies, 3D scenes, and AI datasets that previous low-profile cards lacked the capacity for. 4th generation RT Cores with RTX Mega Geometry add photorealistic ray tracing with the ability to process up to 100× more ray-traced triangles, while DLSS 4 Multi Frame Generation ensures smooth rendering and simulation applications. For architects, industrial designers, and VFX professionals working in compact setups, this is a generational leap available without any infrastructure changes.

NVIDIA GPU

Blackwell generation performance without the need for powerful PSUs

Industrial control systems, media studios, and financial institutions have one thing in common—they need GPU power where traditional cards physically won't fit. The NVIDIA RTX PRO 4000 24GB SFF Blackwell Low Profile, with its 2.7" × 6.6" dimensions and 70W TBP without external power, opens up access to the Blackwell generation for slim tower workstations, edge systems, and densely populated server racks. ISV certification for key professional applications and NVIDIA enterprise drivers guarantee proven stability and optimisation for production deployment—regardless of case size or installation location.

NVIDIA GPU
Nvidia GeForce RTX 40

NVIDIA Blackwell architecture

A breakthrough generation for AI, neural rendering, and scientific computing
Nvidia GeForce RTX 40

5th generation Tensor Cores

770 AI TOPS with FP4 for local AI inference and generative AI applications
Nvidia GeForce RTX 40

24GB GDDR7 ECC

432GB/s for large AI models and complex 3D scenes in an SFF format
Nvidia GeForce RTX 40

New streaming multiprocessors

Neural shaders for next-generation, AI-augmented graphics workflows
Nvidia GeForce RTX 40

4th generation RT cores

RTX Mega Geometry for physically accurate scenes with billions of polygons
Nvidia GeForce RTX 40

4× DisplayPort 2.1b

Up to four monitors at 4K@165Hz or two at 8K@100Hz
Nvidia GeForce RTX 40

Half-height 167mm

Deploy anywhere standard cards physically won't fit
Nvidia GeForce RTX 40

PCIe 5.0 x8

Double the throughput of PCIe Gen 4 for data-intensive AI tasks

Specifications

Chip

Chip manufacturer NVIDIA
Model designation RTX PRO 4000
Graphics processor GB203
Chip architecture Blackwell
Number of stream processors 8,960×
Production technology 4 nm

Performance & usage

Use Professional
Card profile FHHL

Memory

VRAM capacity 24 GB
Memory type GDDR7
Bus width 192-Bit

Frequency

Core frequency 790 MHz
Core clock speed (Boost clock) 1,337 MHz
Memory frequency 18,000 MHz (18 GHz)
Effective memory clock (bandwidth) 18 Gbps

Interface

PC connection PCI Express x16 5.0
Bus interface PCIe 5.0 x8
Monitor connection mini DisplayPort 2.1

Technology

DirectX version 12 Ultimate

Cooling system

Cooling type Active
Cooling size Dual-slot
number of fans 1 ×

Dimensions

Width 167 mm
Depth 69 mm

Consumption

TDP 70 W

AI

Total number of TOPS 770
Supported precision FP64, FP32, BF16, FP16, FP8, INT8

Tops

Total number of TOPS 770

AI specifications

Primary AI use Workstation AI
CUDA version 12.x
Code:  PNYrpb03
Warranty: 36 months
Links: Producer's Website:
Manufacturer information

PNY Technologies Europe, ZAC du Phare 9 rue Joseph Cugnot 33708 Mérignac cedex, France https://www.pny.com/en-eu/support-footer/technical-support

Show more