Nvidia has officially introduced the RTX PRO 5500 Blackwell, a new professional graphics card designed for workstation deployment. The hardware is specifically optimized for teams handling demanding tasks such as AI agent development, physical simulations, and complex rendering workloads.
The card features 21,760 CUDA cores, matching the core count found in the GeForce RTX 5090 series. It is equipped with 84GB of GDDR7 memory that supports ECC. The device offers a memory bandwidth of 1,398 GB/s, utilizing a 448-bit memory bus with an effective speed of 25 Gbps.
In terms of connectivity and power, the RTX PRO 5500 includes three NVENC and three NVDEC decoders, along with up to four DisplayPort 2.1b ports. The card has a maximum power consumption of 600W and is available with both active air cooling and liquid cooling solutions.
