Nvidia continues to fuel the AI boom with the launch of the new RTX Pro 5000 72GB, a high-memory variant built on the Blackwell architecture. While not designed for gaming, this GPU is engineered to handle the most demanding AI workloads with exceptional efficiency. The standout feature is its 72GB of GDDR7 memory—50% more than the standard RTX Pro 5000’s 48GB—making it a powerful tool for training and running large-scale AI models. The RTX Pro 5000 72GB shares the same GB202 silicon as the GeForce RTX 5090 and the RTX Pro 6000, but with 110 of the possible 192 streaming multiprocessors (SMs) enabled, representing about 57% of the die’s potential. This configuration delivers 14,080 CUDA cores, 440 Tensor Cores, and 110 Ray Tracing Cores, all operating at a boost clock of 2,617 MHz. Despite the significant memory upgrade, power consumption remains unchanged at 300 watts, drawing power through a single 16-pin connector. This efficient power profile, combined with a dual-slot blower-style cooler, allows systems to support up to four of these cards—ideal for AI workstations and data centers. The GPU features a 384-bit memory bus and a memory bandwidth of 1,344 GB/s, matching the standard RTX Pro 5000. It also maintains a 96MB L2 cache and 176 render output units. In terms of performance, it delivers 73.69 TFLOPS of FP32 compute power—on par with the base model—though the expanded VRAM significantly improves its ability to manage large models and datasets without bottlenecks. Priced between $4,400 and $7,500 for the standard RTX Pro 5000, the 72GB version is expected to cost 20% to 25% more, likely landing in the $5,300 to $9,400 range. This positions it as a strategic middle ground between the more affordable RTX Pro 5000 and the premium RTX Pro 6000, which can cost up to $10,000. The 72GB model offers a compelling value for organizations needing more memory than the base model but not the full capabilities of the top-tier card. Nvidia has not yet announced a release date or official pricing. However, the card is expected to appeal to enterprises, research labs, and AI developers seeking scalable, high-capacity hardware without the premium price tag of the RTX Pro 6000. With AI demand surging and data-intensive models becoming the norm, this new GPU underscores Nvidia’s continued dominance in the AI accelerator market.