NVIDIA RTX 5090 New Era of Graphics

Discover the revolutionary Blackwell architecture, incredible performance and technologies of the future

0
GB GDDR7
0
CUDA Cores
0
W TGP

RTX 5090: Flagship Performance

Blackwell Architecture

A completely new architecture built on TSMC's 4nm process node, delivering incredible performance and energy efficiency.

🚀

3rd Gen RT Cores

Real-time ray tracing at a whole new level

🧠

5th Gen Tensor Cores

AI and machine learning at revolutionary speed

GDDR7 Memory

32 GB of ultra-fast memory for any task

RT
AI
CUDA
GDDR7

RTX Series Comparison

Model
CUDA Cores
Memory
TGP
Performance
RTX 5090
Flagship
21,760
32 GB GDDR7
575W
100%
RTX 5080
High-End
10,752
16 GB GDDR7
400W
78%
RTX 5070 Ti
Performance
8,960
16 GB GDDR7
300W
65%
RTX 5070
Mainstream
6,144
12 GB GDDR7
250W
52%

Evolution of NVIDIA Architectures

2025

Blackwell (RTX 50)

4nm TSMC
+70% performance

Revolutionary architecture with GDDR7 support and enhanced ray tracing

2022

Ada Lovelace (RTX 40)

5nm TSMC
+50% performance

Significant leap in efficiency and DLSS 3 support

2020

Ampere (RTX 30)

8nm Samsung
+30% performance

First generation with second-generation ray tracing support

2018

Turing (RTX 20)

12nm TSMC
First RT Cores

Beginning of the real-time ray tracing era

Detailed Specifications

🎮

Gaming Performance

4K Ultra: 120+ FPS
8K High: 60+ FPS
VR Ready: 4K per eye
🎨

Creative Tasks

3D Rendering: 3x faster
AI Processing: 5x faster
Video Editing: 8K 60FPS
🔧

Technical Parameters

Process Node: 4nm TSMC
Transistors: 92 billion
Boost Clock: 2850 MHz