Specifications

Overview

The NVIDIA RTX PRO 5500 Blackwell Workstation Edition slots between the RTX PRO 5000 and RTX PRO 6000 to close the memory-capacity gap in NVIDIA's professional Blackwell line-up. It carries 84GB of GDDR7 with ECC — the second NVIDIA part to reach that capacity after the China-exclusive RTX PRO 6000D — achieved with 28 24Gb GDDR7 modules in a clamshell layout across a 448-bit interface running at 25 Gb/s, for 1,398 GB/s of bandwidth. That memory subsystem is shared with the RTX PRO 6000D, while the compute configuration tracks close to the RTX 5090. NVIDIA positions the card for organizations that want to centralize GPU power into IT-managed racks and share it across users: it takes up to 600W, is offered in active air-cooled and liquid-cooled RXM thermal solutions, and drives up to four DisplayPort 2.1b outputs. Sixth-generation NVDEC engines with 4:2:2 H.264/HEVC decode plus AV1 support make it well suited to live media pipelines as well as agentic AI, physical simulation, and graphics workloads.