NVIDIA has redefined its professional graphics lineup by rebranding it as ´RTX PRO´ and launching the RTX PRO 6000, the first workstation GPU built with 96 GB of GDDR7 memory—made possible through innovative 3 GB memory modules. These modules are distributed evenly on both sides of the PCB, with 16 modules per side, bringing unprecedented memory capacity and supporting error-correcting code to ensure high reliability for demanding workloads.
The new PCB design, revealed in a recent Chiphell forum leak, does away with the previous 12 V-6×2 power connector, instead incorporating four solder points to accommodate a cable extension. This modification prepares the card for both Server and Max-Q editions, with the power inputs conveniently shifted to the rear, streamlining the overall footprint. The workstation GPU maintains its full GB202 Blackwell GPU and full memory configuration, balancing innovation with consistent high-end performance.
The RTX PRO 6000 Blackwell series is set to arrive in three primary configurations: Workstation, Server, and Max-Q. The Workstation and Server variants are equipped with 24,064 CUDA cores, 96 GB of GDDR7 ECC memory, and a 600 W power budget—enabling powerful performance for desktop towers and rack-mounted systems. The Max-Q version retains the same high-end GPU and memory capacity but operates at lower clock speeds and a 300 W power limit, targeting compact and noise-sensitive environments without sacrificing computational capabilities. This design positions the RTX PRO 6000 as a versatile and forward-looking tool across professional, scientific, and Artificial Intelligence applications.
