SERVCHIP
Interactive view
ARCHITECTURE
Ada Lovelace
SERIES
RTX 4000
Memory
20GB GDDR6 ECC
Bandwidth
320 GB/s
TDP
130W
In StockRTX 4000·Best for: Compact Data Center AI & Graphics
NVIDIA RTX 4000 Ada Generation
Ada Lovelace
|Data Center GPUsOverview
The NVIDIA RTX 4000 Ada Generation for data center deployments is a compact, power-efficient accelerator featuring 20GB of GDDR6 ECC memory with 320 GB/s bandwidth. Built on the Ada Lovelace architecture with 6,144 CUDA cores, 4th-generation Tensor Cores, and 3rd-generation RT Cores, it excels at edge AI inference, video analytics, and virtual desktop infrastructure. Its single-slot PCIe form factor and low 130W TDP make it ideal for dense server deployments requiring AI acceleration alongside graphics processing.
Key Features
- 20GB GDDR6 ECC memory with 320 GB/s bandwidth
- Ada Lovelace architecture with Tensor Cores
- 6,144 CUDA cores with 3rd-gen RT Cores
- Single-slot PCIe form factor for dense deployments
- Hardware AV1 encode and decode
Ideal Use Cases
ai inferenceedge computingvirtualization
Full Specifications
Complete technical specifications for the RTX 4000.
Memory
- Memory
- 20GB GDDR6 ECC
- Memory Bandwidth
- 320 GB/s
Compute
- Tensor Cores
- 192 (4th-gen)
- CUDA Cores
- 6144
- TF32 TFLOPS
- 26.7
Interconnect
- Interconnect
- PCIe 4.0 x16
Physical
- TDP
- 130W
- Form Factor
- PCIe 4.0 single-slot
- Cooling
- Active Fan
Manufacturing
- Launch Date
- Q1 2024
- Manufacturing Process
- 5nm
