Tuesday, October 21st 2025

NVIDIA RTX PRO 5000 "Blackwell" GPU with 72 GB GDDR7 Memory Appears

NVIDIA's RTX PRO 5000 "Blackwell" GPU has officially appeared with 72 GB of GDDR7 ECC memory, on the company's product page. Initially, NVIDIA launched an RTX PRO 5000 Blackwell GPU with 48 GB of GDDR7 memory spread across 24 modules with the capacity of 2 GB, sandwiching the PCB on both sides. However, NVIDIA has now decided to update the GPU with the new 3 GB memory modules, which are upping the total capacity to 72 GB on a 512-bit bus with 1,344 GB/s memory bandwidth, making the card a much more attractive offer for professionals working in the fields of data science, AI, HPC, and many other areas such as professional video editing. However, this 24 module design with 12 GDDR7 modules on each side is less impressive than the RTX PRO 6000 Blackwell GPU, which arranges 16 such modules on each side of the PCB to achieve a remarkable 96 GB capacity.

As a reminder, the RTX PRO 5000 GPU is based on the GB202 GPU and carries 14,080 CUDA cores, 440 TMUs, and 176 ROPs. Upgraded NVENC/NVDEC blocks speed up high-quality encode/decode for live production and fast video editing. Multi-Instance GPU (MIG) gives IT and cloud/VDI admins an easy way to split a GPU into isolated instances, so more users get guaranteed, accelerated performance. NVIDIA has wrapped all of this in a 300 W TDP in a dual-slot GPU, with a blower fan. Pricing structure is still unknown, but given that its older sibling, the RTX PRO 6000 Blackwell, is retailing just under $10,000, we could see the 72 GB version just a grand or two lower. Additionally, as we see NVIDIA incorporating 3 GB GDDR7 memory modules inside more products, the company could be gearing up for a "SUPER" overhaul of the Blackwell gaming lineup.
Add your own comment

12 Comments on NVIDIA RTX PRO 5000 "Blackwell" GPU with 72 GB GDDR7 Memory Appears

#1
jabbadap
How can you get 72GB on 512bit bus(72GB/16=4.5)? Is there an error on specs or some mix maxed memory chips on both side. 48GB is fine with 16 times 3GB gddr7 chips single sided.

Edit: Most probably it has 384 bit bus with 12 chips per side, that adds up.
Posted on Reply
#2
dismuter
jabbadapHow can you get 72GB on 512bit bus(72GB/16=4.5)? Is there an error on specs or some mix maxed memory chips on both side. 48GB is fine with 16 times 3GB gddr7 chips single sided.

Edit: Most probably it has 384 bit bus with 12 chips per side, that adds up.
Yes it's a mistake in the article. The spec sheet on TPU's own website says it has a 384-bit bus.
Posted on Reply
#3
jabbadap
dismuterYes it's a mistake in the article. The spec sheet on TPU's own website says it has a 384-bit bus.
That spec sheet is directly from nvidia, so probably just copy paste error on that side.
Posted on Reply
#4
dismuter
jabbadapThat spec sheet is directly from nvidia, so probably just copy paste error on that side.
I see, so NVIDIA can't even get their own specs right...
Posted on Reply
#5
Tomorrow
This is essentially what 5080 Ti would be (with the exception of the 72GB capacity) if Nvidia ever decided to release it.
Posted on Reply
#6
Xaser04
The underlying GPU feels like what the 5080 should have been (on the lower end vs core count of the 5090) with ~24GB VRAM rather than the anemic thing we ended up with.
Posted on Reply
#7
wesley1
More interesting is total board power 300W at 65 TFLOPS, and huge memory.

That's a card faster from upcoming 5080 Super at same architecture (Backwell). Nvidia sells garbage to individual clients.
Posted on Reply
#8
Nater
They are all so expensive and in short supply, the Blackwell workstation cards are essentially paper launches.
Posted on Reply
#9
bonehead123
Well, my company is already rocking the Pro 6000's (~135.72 of them), so this PR is a big yawn for them, but I'm sure there are plenty of folks who don't need all that VRAM & performance, and will probably buy these due to the lower prices :D
Posted on Reply
#10
RootinTootinPootin
wesley1More interesting is total board power 300W at 65 TFLOPS, and huge memory.

That's a card faster from upcoming 5080 Super at same architecture (Backwell). Nvidia sells garbage to individual clients.
we're all technically using the garbage, even the 5090 is a garbage RTX 6000 bin with tight limits..
Posted on Reply
#12
VoidRunner96
wesley1More interesting is total board power 300W at 65 TFLOPS, and huge memory.

That's a card faster from upcoming 5080 Super at same architecture (Backwell). Nvidia sells garbage to individual clients.
Who would you cater to more... one who gives you lets say, +$250 profit or +$5k (just rando numbers) but you get the picture. Nvidia was already making alot outside of gaming, as they where in autos also, just to name one.

Garbage is Kinda harsh lol, but i get ya
Posted on Reply
Add your own comment
Nov 14th, 2026 18:58 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts