Architecture & Performance
Built on NVIDIA Blackwell architecture (revision 2.0) with 10,752 CUDA cores for strong compute, ray tracing, and AI throughput
Clock Speeds: Base ~2,295 MHz, Boost up to ~2,640 MHz in factory-overclocked "OC" trim
AI Features: DLSS 4 with multi-frame generation, accelerated by 5th‑gen Tensor Cores and Ray Reconstruction support
Memory & Bandwidth
VRAM: 16 GB of ultra-fast GDDR7 at 30 Gbps over a 256‑bit interface, yielding ~960 GB/s bandwidth
Cooling, Power & Thermals
Power Draw: ~360 W TDP, requiring a single 16‑pin PCIe 5.0 connector; recommended PSU is 750 W+
Cooling System: Triple‑fan (axial fans) with vapor chamber heatsink (VC design) plus idle-fan‑stop functionality for quieter operation
Thermal Limit: Maximum safe temperature around 88 °C
Connectivity & Physical Attributes
Display Interfaces: 3 × DisplayPort 2.1b + 1 × HDMI 2.1b, supporting up to four simultaneous displays
Max Resolution: Up to 7680×4320 (8K) output via DSC support
Dimensions: Approximately 312 mm × 127 mm × 66 mm, occupying about 3.5 PCIe slots
