Friday, October 24th 2025

AMD Radeon AI PRO R9700 GPU Arrives October 27 at $1,299 for Retail

AMD has finally set the retailer date and the price tag for its professional Radeon AI PRO 9000 series graphics cards, suitable for tasks like professional visualization or workstations that can accommodate up to four cards. Set to launch on October 27 for retailers, the flagship Radeon AI PRO R9700 will arrive at $1,299. Previously, AMD focused solely on the OEM and system integrator segment, leaving retailers and individuals building their own workstations for AI and professional visualization workloads without options. The company aimed to launch this GPU for retailers by the end of July, but supply constraints prevented this, as OEMs and system integrators consumed the entire supply.

These Radeon AI PRO R9700 GPU features a 4 nm "Navi 48" GPU, equipped with 64 RDNA 4 compute units, providing 4,096 stream processors and 128 AI accelerators for enhanced matrix operations across various data formats. A key improvement over the RX 9070 XT gaming GPU is the increased memory capacity, now doubled to 32 GB. The memory subsystem includes 20 Gbps GDDR6 across a 256-bit wide interface, delivering 640 GB/s of memory bandwidth, supported by a 64 MB 3rd Gen Infinity Cache. The AMD Radeon AI PRO R9700 achieves up to 191 TeraFLOPS FP16 dense performance and up to 1531 TOPS INT4 sparse performance, all within a power footprint of up to 300 W. The card is packed in dual-slot design with blower-style cooler, so dense multi-GPU configurations can be easily achieved.
Source: via VideoCardz
Add your own comment

56 Comments on AMD Radeon AI PRO R9700 GPU Arrives October 27 at $1,299 for Retail

#51
LabRat 891
TomorrowYeah without explicit support in the specs i would assume no. Same as ECC only in Linux. I assume it's some kind of software based ECC. If it were hardware then it would be OS agnostic. Shame. AMD could have put more work into this card. Any idea about SR-IOV ?
ECC is somehow wrapped-in with ROCm. So yeah, software ECC is likely [I don't feel like yanking apart my card just to read the chip's markings]
TBF on every example I've seen or read about of the R9700, all use cool-running Samsung VRAM. Both examples I've played with seem to do 2666MHz(2680 set) easy-peasy.

SR-IOV is a hard no. 'Knew that one going into it, after mucking about w/ MI25s/WX9100s awhile back. Not even the W7900 has SR-IOV, IIRC.
AFAIK, only the Radeon Instinct cards (and, at one brief point in time, the VII Pro) have/had (pedantically circumstantial) SR-IOV support.
TomorrowAMD could have put more work into this card.
Disagree.
This is a high-margin product for a (possibly briefly capturable) market. Things like CAD certification/PRO Driver support, hardware ECC VRAM, SR-IOV support, etc. are all exclusive to even-higher margin products, and could increase per-unit costs.
Posted on Reply
#52
csendesmark
If you run your AI models at home this is not a bad option!
Posted on Reply
#53
Vayra86
W1zzardI doubt it, why would anyone sample it based on the below.

But here's my mini review: Gaming: no difference, Price: more expensive, Compute: no CUDA support, but good when supported in the software you use, good for edge cases that hit that 16 .. 32 GB VRAM workload window, or you control the whole software stack or can whip your devs to port, and they are cheaper than buying NV (or you are limited by NV being sold out for months)

If a success, bad for gamers, because less GPUs available for them and AMD will focus even more on AI sales
This is why I can't ever leave TPU. The sheer amount of common sense in a world so devoid of it. Its just too good, even when its so simple.
BlueberriesTo your point, no, there are games that benefit from more than 16GB of memory, especially at high resolutions with Ray Tracing, "Supersampling," or Frame Generation enabled. All of these eat up memory and you can utilize a much larger buffer.

That being said, this isn't a gaming card.
Nyeah... if you crank up the res high enough and in case you could possibly use some sort of mGPU solution you might be able to scale up and beyond the current capability of the regular graphics card stack.

But that's really just an exercise in pointlessness, if not insanity, in this day and age.
Posted on Reply
#54
Tomorrow
Vayra86This is why I can't ever leave TPU.
It's the excellent forum software for me. The dynamic notification if someone has posted while im typing, the multi quote insert and other features.
A far cry from some other tech forums stuck on old phpbb2 and still refusing to upgrade.
Posted on Reply
#55
Rover4444
csendesmarkIf you run your AI models at home this is not a bad option!
Only for LLMs. If you're doing literally anything else AI-wise then NVIDIA is king.
LabRat 891Disagree.
This is a high-margin product for a (possibly briefly capturable) market. Things like CAD certification/PRO Driver support, hardware ECC VRAM, SR-IOV support, etc. are all exclusive to even-higher margin products, and could increase per-unit costs.
LMAO, then maybe AMD should've done that and have it be even more compelling. For imagegen AMD is trash and for textgen Vulkan is at parity with ROCm, and ROCm 7 was a total sham for RDNA. Even the B50 is certified, has ECC memory, AND has SR-IOV support, so what's AMD's excuse now?
Posted on Reply
#56
csendesmark
Rover4444Only for LLMs. If you're doing literally anything else AI-wise then NVIDIA is king.
I agree, You need CUDA for most of the other things.
Posted on Reply
Add your own comment
Nov 20th, 2026 17:07 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts