The RTX Pro 6000 is Nvidia's brand-new flagship graphics card for professional users. The GPU was primarily developed for AI applications, game developers and other professional users who need lots of video memory. While the Nvidia GeForce RTX 5090 is equipped with "only" 32GB GDDR7 VRAM, the desktop and server versions of the RTX Pro 6000 offers a full 96GB of GDDR7 with a bandwidth of 1.6 TB/s.
With 96GB of VRAM, the RTX Pro 6000 competes against AMD’s Ryzen Strix Halo in terms of AI workloads, and the graphics card should be able to process equally large AI models at a significantly higher speed. The GPU has a TDP between 400 and 600 watts, and it supports both PCIe 5.0 and DisplayPort 2.1. The slimmer Max-Q variant may be particularly intriguing for users who are planning to put multiple graphics cards in a single PC case.
The Nvidia RTX Pro 6000 is also offered as a laptop GPU, but that model is limited to just 24GB VRAM, similar to the GeForce RTX 5090 Laptop. With the RTX Pro 3000, 2000, 1000 and 500, Nvidia also offers more affordable professional GPUs based on the Blackwell architecture. Nvidia has not yet provided any details regarding the number of CUDA cores or clock speeds of the new RTX Pro graphics cards.
Price and availability
Nvidia has yet to reveal the official MSRPs of its new professional graphics cards. The RTX Pro 6000 is expected to start shipping in April, and pre-built systems from Dell, HP and Lenovo are expected to hit the market in May.