NVIDIA introduces RTX PRO 6000 'Blackwell' GPU series: 24064 cores, 96GB memory and up to 600W

NVIDIA has just unveiled its new RTX PRO 6000 "Blackwell" graphics card, based on the same GB202 GPU and GDDR7 memory that the RTX 5090 uses, but with a whopping 96GB of GDDR7 memory.

5

The new NVIDIA RTX PRO 6000 Blackwell GPU features the GB202 GPU with 24,064 cores inside of 188 SMs, compared to the 21,760 cores and 170 SMs inside of the same GB202 GPU in the GeForce RTX 5090. The full-fat GB202 GPU features 192 SMs, so we're not even maxed out here with the new RTX PRO 6000.

We do have the full 600W of TDP support through a single 12V-2x6 power connector, offering 25W more TDP than the 575W on the RTX 5090. With all of this in tow, NVIDIA's new RTX PRO 6000 Blackwell GPU features 4000 AI TOPs, 125 TFLOPS of FP32 compute performance, and 380 TFLOPS of RT performance (a 19% increase over the RTX 5090 in all performance figures).

Not only do we have a beefier GB202 GPU, but that 19% increase on GPU specs over the RTX 5090 are far more visible on the RTX PRO 6000 Blackwell GPU when it comes to memory capacity. NVIDIA's consumer-focused GeForce RTX 5090 features 32GB of GDDR7 memory, but the new RTX PRO 6000 Blackwell GPU has a far larger 96GB pool of GDDR7 memory: a 3x increase.

We have the same 512-bit memory bus and same 28Gbps GDDR7 memory modules, with 1792GB/sec of memory bandwidth in total. NVIDIA is using the same double-flow-through design and dual-slot form factor, supporting MIG (Multi-Instance-GPU) capability with 4 x 24GB, 2 x 48GB, or 1 x 96GB modes.

Our friends over at Wccftech have obtained some new benchmark figures using GameTechBench, with the new benchmark (available on Steam) featuring full path tracing (CGI rendering) and provides a good look into the more powerful GB202 GPU + 96GB of GDDR7 memory

5

We have a 5% performance lead at 4K in the GameTechBench over the RTX 5090, with the RTX PRO 6000 GPU just behind in 1440p real-time ray tracing tests. In the offline Path Tracing rendering mode, the new NVIDIA RTX PRO 6000 Blackwell GPU is only 2% faster than the RTX 5090.

It will be interesting to see if we have 19% performance uplifts in full AI workloads, especially with that far larger 96GB pool of GDDR7 memory, a 3x increase in VRAM capacity over the 32GB on the RTX 5090.