NVIDIA Unveils RTX PRO 6000 Series Graphics Cards with 96GB of Memory

NVIDIA has announced its new RTX PRO 6000 Blackwell Series lineup. Built on the Blackwell architecture, these innovations promise to transform workflows in artificial intelligence, 3D modeling, and scientific computing with unprecedented performance and cutting-edge solutions.

The flagship RTX PRO 6000 Workstation Edition sets a new benchmark by combining 24,064 CUDA cores, 768 Tensor cores, and 192 fourth-generation RT cores. It features 96GB of error-correcting GDDR7 memory on a 512-bit bus, delivering unmatched data processing speeds for rendering, simulations, and neural network training. With a 600W power draw, the card utilizes a PCIe 5.0 interface and supports resolutions up to 8K at 165Hz via DisplayPort 2.1.

The Blackwell Series is available in three configurations:

For the first time in its professional lineup, NVIDIA introduces mobile accelerators—ranging from the RTX PRO 5000 down to the RTX PRO 500 Blackwell. Notebook versions come equipped with up to 24GB of GDDR7 memory and up to 10,496 CUDA cores, ensuring uncompromised mobility. For less demanding tasks, the RTX PRO 4500 and 4000 models are available with 24–32GB of memory.

Desktop RTX PRO 6000 cards will hit the market in April via PNY and TD SYNNEX, with systems from BOXX, Dell, and HP following in May. The lower-tier RTX PRO 5000/4500/4000 models are expected to roll out in the summer, while notebook solutions will be offered by Dell, Lenovo, and Razer by the end of 2025. Pricing has not yet been disclosed, though experts anticipate a premium positioning.