NVIDIA RTX 6000 X Pro Blackwell with 96GB and 600W: Extreme Power for AI and Creativity | GPU hardware | Types of computer hardware | List of hardware components | Turtles AI

NVIDIA is about to launch the RTX PRO 6000 X Blackwell workstation, a GPU with an advanced architecture and high-end specifications. Equipped with 96 GB of GDDR7 memory and a 600 W TBP, this card is designed for AI applications and content creation. A variant with 48 GB on a 384-bit bus is also expected.

Key points:

• Blackwell Architecture: The new GPU uses the GB202-870 architecture for high performance.
• 96 GB of GDDR7 memory: An ideal amount of VRAM for AI and intensive workloads.
• 600 W TBP: Higher power consumption than the RTX 5090 to maximize performance.
• 48 GB variant: Less powerful version with a reduced memory bus of 384 bits.

NVIDIA is all set to introduce the new RTX PRO 6000 X Blackwell, a high-end workstation graphics card that promises top-notch performance thanks to a design based on the GB202-870 GPU and equipped with a massive amount of 96 GB of GDDR7 memory. Intended for areas such as AI and content creation, this GPU stands out for its advanced architecture and a hardware configuration designed to satisfy the most demanding workloads. The TBP of 600 W, slightly higher than that of the GeForce RTX 5090, suggests a high power consumption, but necessary to better support the memory modules and ensure operational stability in the most demanding applications. The card has been spotted in NBD shipping logs, with units headed to India for validation tests, signaling its imminent market launch.

In parallel, NVIDIA also seems to be working on a less powerful version, the RTX PRO 6000, without the "X" suffix and featuring a configuration with 48 GB of GDDR7 memory on a reduced 384-bit bus. This solution seems interesting since at the moment no other GPU of the RTX 50 series is known to have a similar memory interface. While the RTX PRO 6000 X uses the same architecture as the RTX 5090, the 48 GB model could use a different chip, whose identity remains a mystery.

The comparison with the RTX 5090 highlights some significant differences: although both cards adopt a 512-bit bus, the workstation PRO 6000 X triples the amount of VRAM, making it ideal for scenarios where memory capacity is a critical factor. The choice of GDDR7 memory represents a further step forward compared to the previous generation, with an increase in bandwidth and greater efficiency. The 25W TBP increase over the 5090, while not revolutionary, demonstrates NVIDIA’s intent to maximize performance, making the most of the card’s potential.

The company’s strategy therefore seems to be oriented towards diversifying the offering in the professional segment, expanding the range of solutions based on Blackwell. The RTX PRO 6000 X is configured as a flagship card, capable of handling extremely heavy workloads, while the standard version could represent a more accessible option for those who need high computing power, but without excess.

With the sighting in the shipping documents, the official debut of these new GPUs could be imminent, marking a further step forward in the evolution of NVIDIA’s graphics solutions.