- Packed with 24 gb of ultra fast gddr7 memory and ecc error correction, this gpu maintains data integrity during intensive ai training, large scale simulations and high resolution rendering, reducing costly crashes and re renders.
- Designed as a single slot, 140 w blower cooled card, it fits easily into dense rack mount or compact workstations while keeping power draw and heat low, enabling multi gpu configurations without overloading the chassis.
- Supports pcie 5.0 x16 and four displayport 2.1b outputs, delivering up to dual 8k @ 100 hz or four 4k @ 165 hz, plus hardware accelerated nvenc/av1 encoding, so creators can stream, edit and visualize content at studio grade performance.
Description
Accelerate your professional engineering, data science, and AI deployment with the NVIDIA RTX PRO 4000 Blackwell 24GB Workstation Graphics Card. Built on the revolutionary NVIDIA Blackwell architecture, this ultra-efficient powerhouse delivers groundbreaking neural graphics and enterprise-grade performance within a compact footprint. Equipped with a massive 24GB of ultra-fast GDDR7 memory with Error-Correcting Code (ECC), it guarantees extreme stability for mission-critical tasks, local LLM prototyping, and rendering highly intricate 3D assets.
Featuring an incredibly slim, ultra-efficient single-slot design and drawing a maximum of just 140W, it easily drops into dense multi-GPU rackmounts or compact office workstations. Armed with fifth-generation Tensor Cores supporting ultra-efficient FP4 precision and fourth-generation RT Cores for complex ray-traced geometries, it offers unmatched compute densities. Complete with a cutting-edge PCI Express 5.0 x16 interface and four DisplayPort 2.1b outputs capable of driving up to dual 8K displays at 100Hz, this professional GPU is fully ISV-certified to streamline your design, AI, and generative simulation pipelines without compromise.
Specifications
Technical Parameter
Details
Model / MPN
NVIDIA RTX PRO(TM) 4000 Blackwell (900-5G147-2270-000)
GPU Architecture
NVIDIA Blackwell
CUDA Cores
8960 Units
Tensor Cores
5th Generation (Supports FP4 Precision & DLSS 4)
Ray Tracing Cores
4th Generation (RTX Mega Geometry Support)
AI Performance
Up to 1178 AI TOPS
Memory Size / Type
24GB GDDR7
Memory Reliability
Error-Correcting Code (ECC) Support
Memory Interface
192-bit
Memory Bandwidth
672 GB/s
Bus Interface
PCI Express 5.0 x16
Display Connectors
4x DisplayPort 2.1b
Max Resolution / Displays
Up to 4x 4K @ 165Hz or 2x 8K @ 100Hz
Video Hardware Acceleration
2x 9th-Gen NVENC (AV1, H.264/HEVC 4:2:2), 2x 6th-Gen NVDEC
Max Power Consumption
140 W
Power Input
1x PCIe CEM5 16-pin (12VHPWR/12V-2x6 style)
Cooling Architecture
Blower Active Fan (Optimized for Multi-GPU Layouts)
Form Factor
4.4" H x 9.5" L, Full Height, Single-Slot
Enterprise Ecosystem
ISV Certified Drivers, NVIDIA RTX Desktop Manager, Frame Lock Support