
NVIDIA is set to redefine workstation graphics with the anticipation surrounding its RTX PRO 6000 “Blackwell”GPU. With a staggering configuration exceeding 24, 000 CUDA cores and an impressive 96 GB of VRAM, this graphics card promises to deliver unparalleled performance for professional applications.
NVIDIA Unveils the Power of Blackwell: The RTX PRO 6000 Leads the Charge
Recent reports reveal that NVIDIA is actively developing its next-generation Blackwell workstation graphics card, falling under the “RTX PRO”series. At the forefront is the flagship model, the RTX PRO 6000, which has now had its specifications confirmed after some preliminary data was revealed.
A critical insight came from a recent listing by Leadtek, which disclosed the initial specifications of this formidable GPU. For starters, the Blackwell architecture will feature 188 streaming multiprocessors (SMs) out of a potential 192 SMs on the GB202 GPU. While this chip is not fully enabled, it still significantly surpasses the gaming-centric RTX 5090.
NVIDIA RTX PRO™ 6000 Blackwell Workstation Edition – Blackwell GPU – 24, 064 CUDA Cores – 752 Tensor Cores – 188 RT Cores – 96GB DDR7 Memory with ECC – 600W https://t.co/MhxxgDzSc5 pic.twitter.com/ss07xVeyXW
— Posiposi (@harukaze5719) March 8, 2025
The NVIDIA RTX PRO 6000 “Blackwell”is set to house 24, 064 cores, representing a 10.5% increase over the 21, 760 cores found in the RTX 5090. In addition to the core improvements, this GPU will incorporate an enhanced configuration with 752 tensor cores and 188 ray tracing (RT) cores. However, what stands out the most is its extraordinary memory capacity.
Unlike the NVIDIA GeForce RTX 5090, which is equipped with 32 GB of GDDR7 memory on a 512-bit bus, the RTX PRO 6000 “Blackwell”will boast a remarkable 96 GB of GDDR7 (ECC) memory utilizing a 384-bit bus. It’s important to note that a 384-bit bus is required to achieve a memory capacity of 96 GB; a 512-bit configuration results in maximum capacities of 32, 64, or 128 GB. To facilitate this memory capacity, NVIDIA will deploy 24 memory modules, each with a 4 GB capacity. While speed specifics remain unconfirmed, projections suggest speeds will be no less than 28 Gbps.

Furthermore, the total board power (TBP) of the NVIDIA RTX PRO 6000 is expected to be around 600W. This figure aligns with the maximum specifications permissible through a single 12V-2×6 16-pin power interface. Adequate thermal management for such a powerful card necessitates extensive design efforts, and it is anticipated that NVIDIA’s thermal engineering team has engineered a new reference design, potentially featuring either a single blower-type or dual-axial fan configuration. The latest listings indicate that a launch is imminent, so more detailed information can be expected in the coming weeks. Price projections for this workstation powerhouse range between $10, 000 and $15, 000, or potentially even higher.
NVIDIA Workstation Graphics Card Comparison
Graphics Card | RTX PRO 6000 X | RTX 6000 Is Here | RTX A6000 | Quadro RTX 8000 | Quadro RTX 6000 | GV100 Frame |
---|---|---|---|---|---|---|
GPU | Blackwell GPU | Ada Lovelace GPU | Ampere GPU | Turing GPU | Turing GPU | GPU Vault |
GPU WeU | GB202 | AD102 | GA102 | TU102 | TU102 | GV100 |
Process Node | 5nm | 5nm | 8nm | 12nm | 12nm | 12nm |
Die Size | 750mm² | 608mm² | 628mm² | 754mm² | 754mm² | 815mm² |
GPU Cores | TBD | 18, 176 Cores | 10, 752 Cores | 4, 608 Cores | 4, 608 Cores | 5, 120 Cores |
Tensor Cores | TBD | 568 Cores | 656 Cores | 576 Cores | 576 Cores | 640 Cores |
Boost Clock | TBD | 2.50 GHz | 1.80 GHz | 1.77 GHz | 1.77 GHz | 1.62 GHz |
Single Precision | TBD | 91.1 TFLOPs | 38.7 TFLOPs | 16.31 TFLOPs | 16.31 TFLOPs | 16.66 TFLOPs |
Ray Tracing Performance | TBD | 210.6 TFLOPs | 75.4 TFLOPs | 10 GigaRays/Sec | 10 GigaRays/Sec | N/A |
Memory | 96 GB GDDR7? | 48GB GDDR6 | 48GB GDDR6 | 48GB GDDR6 | 24GB GDDR6 | 32 GB HBM2 |
NVLINK VRAM | N/A | N/A | 96 GB with NVLINK | 96 GB with NVLINK | 48 GB with NVLINK | N/A |
Memory Bus Width | TBD | 384-bit | 384-bit | 384-bit | 384-bit | 4096-bit |
Memory Bandwidth | TBD | 960GB/s | 768GB/s | 672 GB/s | 672 GB/s | 870GB/s |
Thermal Design Power (TDP) | 600W? | 300W | 300W | ~225W | ~200W | 250W |
Estimated Launch Price | TBD | $6, 800 | $4, 650 | $10, 000 | $6, 300 | $9, 000 |
Expected Launch Date | 2025 | Q1 2023 | Q4 2020 | Q4 2018 | Q4 2018 | 2018 |
For further details, refer to sources: Harukaze5719, Mark Brown.
Leave a Reply ▼