ASUS ROG ASTRAL LC GEFORCE RTX 5090 OC EDITION 32GB -GRAPHICS CARD

The ASUS ROG Astral LC GeForce RTX 5090 OC Edition 32GB is ASUS's top-of-the-line liquid-cooled graphics card, based on the NVIDIA Blackwell architecture. The Astral LC model combines powerful 360mm AIO liquid cooling, high factory overclocking, and premium-grade components to deliver exceptional performance for demanding 4K and 8K gaming, AI computing, and professional content creation.

Technical Specifications

Graphics Processor: NVIDIA GeForce RTX 5090
Architecture: NVIDIA Blackwell
Video Memory: 32 GB GDDR7
CUDA Cores: 21,760
Memory Bus: 512-bit
Memory Speed: 28 Gbps
Memory Bandwidth: 1,792 GB/s
Interface: PCI Express 5.0 x16
Tensor Cores: 5th Gen
RT Cores: 4th Gen
Cooling: ROG Astral LC AIO Liquid Cooling
Radiator: 360 mm
Fans: 3 × 120 mm radiator fans + integrated auxiliary fan on the card
Cold Plate: Full-coverage copper cold plate
Dual BIOS: Performance / Quiet
RGB Lighting: Aura Sync RGB
Display Outputs: 3 × DisplayPort 2.1b, 2 × HDMI 2.1b
Maximum Resolution: 7680 × 4320 (8K)
DirectX Support: DirectX 12 Ultimate
OpenGL Support: OpenGL 4.6
DLSS Support: NVIDIA DLSS 4
NVIDIA Reflex 2: Yes
NVIDIA Studio: Yes
Power Connector: 1 × 16-pin (12V-2x6)
Total Graphics Power (TGP): approx. 575 W
Recommended Power Supply: 1000 W or more powerful
Power Components: 80 A MOSFET solution for high overclocking potential
Structural Reinforcement: Metal frame structure, protected PCB, and premium ROG components

The ASUS ROG Astral LC GeForce RTX 5090 OC Edition 32GB offers among the most powerful graphics performance on the market combined with highly efficient liquid cooling. Its 32 GB GDDR7 memory, DLSS 4, advanced ray tracing, full-coverage copper cold plate, and 360mm AIO cooling make it an excellent choice for users who want maximum performance, low temperatures, and quiet operation without a separate custom liquid cooling system. It is particularly suitable for high-end gaming, AI development, 3D rendering, video editing, and other extremely demanding workloads.