Tuesday, November 18th 2025

ASUS GeForce RTX 5090 ROG MATRIX Priced at $3,999, Only 1,000 Units Available

ASUS has officially listed its GeForce RTX 5090 ROG MATRIX 30th anniversary limited-edition GPU for $3,999 on its U.S. store. Limited to 1,000 units, this is ASUS's most powerful and expensive card for the RTX 50-series "Blackwell" generation. The build quality justifies the price, as ASUS uses 3 oz copper PCB layers to enhance heat dissipation and minimize voltage drop across components, resulting in improved overall performance and potentially higher overclocking capabilities. The GPU features four large fans, similar to ASUS's RTX 5090 Astral GPU, with three fans in the front and one in the back, utilizing a push-pull air cooling design. The GPU shroud has a prominent ROG red design with a distinct ring on one fan, featuring an acrylic plate with the ROG MATRIX logo.

ASUS also offers an option for BTF-supplied power delivery, along with a 16-pin 12V-2x6 connector, allowing for a total power setting of up to 800 watts, which significantly exceeds the regular 575 watts of the Founders Edition GPU. ASUS claims this supplies a 10% performance boost on its own, and the boost clock frequency is confirmed at 2,730 MHz regular, and 2,760 MHz in OC BIOS mode. Embedded within the GPU is a miniature Bosch Sensortec BMI323 inertial measurement unit (IMU), which combines an accelerometer and gyroscope in a single component. ASUS integrates this sensor into the PCB to continuously monitor whether the graphics card remains level with the motherboard, triggering an immediate alert if the card shifts even slightly downward. More features are listed on ASUS website.
Source: ASUS
Add your own comment

72 Comments on ASUS GeForce RTX 5090 ROG MATRIX Priced at $3,999, Only 1,000 Units Available

#51
CornyOC
R-T-BI kind of doubt that part tbh. PTM doesn't cost THAT much, and LM is expensive.
PTM is singular use and a size large enough for a 5090 die would cost about as much as the most expensive 1g LM tube, which is enough for 2-3 uses.
Posted on Reply
#52
Chrispy_
Bit late for a 5090 launch, isn't it? The 60-series leaks are already happening and there's a non-dismissible amount of info that says the 5090 is being replaced in Sept 2026.

If that's true, Nvidia will likely discontinue the 5090 a few months before that, so this really seems like a desperate last-ditch cash-grab. If someone could afford a 5090 and wanted a 5090 they'd have bought them ages ago.
Posted on Reply
#53
mrnagant
The copper is enough to justify the price!? LOL

What is the spot price of copper again???
Posted on Reply
#54
CornyOC
Chrispy_Bit late for a 5090 launch, isn't it? The 60-series leaks are already happening and there's a non-dismissible amount of info that says the 5090 is being replaced in Sept 2026.

If that's true, Nvidia will likely discontinue the 5090 a few months before that, so this really seems like a desperate last-ditch cash-grab. If someone could afford a 5090 and wanted a 5090 they'd have bought them ages ago.
I heard about 2027 q1 the earliest for 60xx RTX launch, I think 2026 q3 would be for 50xx super launch. I think gaming GPU production is slowing down, instead of 2 years, now it's rather closer to 2.5 years.
Posted on Reply
#55
R-T-B
CornyOCPTM is singular use and a size large enough for a 5090 die would cost about as much as the most expensive 1g LM tube, which is enough for 2-3 uses.
Not for a large OEM. We aren't talking PTM7950 upmarked huge on whatever marketplace because it is not a consumer product. They can buy it at actual cost as a company, or even make their own clone for cheaper.

You? Honestly, we get robbed. PTMs actual build cost is a fraction of what we pay.
Posted on Reply
#56
Chrispy_
CornyOCI heard about 2027 q1 the earliest for 60xx RTX launch, I think 2026 q3 would be for 50xx super launch. I think gaming GPU production is slowing down, instead of 2 years, now it's rather closer to 2.5 years.
5090 launched a few months earlier than everything else in the 50-series product stack and it targeted creators rather than gamers, even though the wealthiest gamers pre-ordered one even before reviews were out.

Nvidia are saying Rubin CPX is slated for Sept 2026, and what they failed to hide on the slides was the fact that Rubin CPX has ROPs and a display block and all the things a gaming GPU needs that AI accelerators don't, so they're releasing something that bears a strong resemblance to an RTX 5090 successor in September 2026. What they call it is anyone's guess at this point. 60-series gaming-focused cards are likely to be 2027 q1 like you are suggesting.
Posted on Reply
#57
CornyOC
Chrispy_5090 launched a few months earlier than everything else in the 50-series product stack and it targeted creators rather than gamers, even though the wealthiest gamers pre-ordered one even before reviews were out.

Nvidia are saying Rubin CPX is slated for Sept 2026, and what they failed to hide on the slides was the fact that Rubin CPX has ROPs and a display block and all the things a gaming GPU needs that AI accelerators don't, so they're releasing something that bears a strong resemblance to an RTX 5090 successor in September 2026. What they call it is anyone's guess at this point. 60-series gaming-focused cards are likely to be 2027 q1 like you are suggesting.
What will that card be then, 5090Ti or a Pro card?
Posted on Reply
#58
Vya Domus
Unironically looks like something AMD would have made during the 2009-2011 era, only thing missing is Ruby.
Posted on Reply
#59
Chrispy_
CornyOCWhat will that card be then, 5090Ti or a Pro card?
Rubin is next gen architecture name, so it'll be the same silicon as the 6090 by the sounds of it. Whether it's the actual 6090 or a new model range for prosumer AI use is anyone's guess.

The 3090 and 4090 were definitely marketed by Nvidia as a premium gaming card that was also good for work.
The 5090 was absolutely not marketed as a gaming card. Even though it's obviously that, Nvidia's series of pre-launch and launch events were selling the 5080 as the flagship for gamers and the 5090 as a workstation showcase for all their AI stuff. Remember Nvidia is an AI company now that still happens to do some graphics on the side....
Posted on Reply
#60
wNotyarD
Chrispy_Rubin is next gen architecture name, so it'll be the same silicon as the 6090 by the sounds of it. Whether it's the actual 6090 or a new model range for prosumer AI use is anyone's guess.

The 3090 and 4090 were definitely marketed by Nvidia as a premium gaming card that was also good for work.
The 5090 was absolutely not marketed as a gaming card. Even though it's obviously that, Nvidia's series of pre-launch and launch events were selling the 5080 as the flagship for gamers and the 5090 as a workstation showcase for all their AI stuff. Remember Nvidia is an AI company now that still happens to do some graphics on the side....
Why not bring back the Titan name?
Posted on Reply
#61
Chrispy_
wNotyarDWhy not bring back the Titan name?
Why not. But it MUST have AI in the name. Google, Microsoft, Meta, AMD, Intel, AMD, and Qualcomm already have products with AI in the model name, so this will need to have at least TWO to make it clear that it's an AI product first and foremost.

Nvidia Titan Rubin AI 8000 Pro (RTX AI Edition)
Posted on Reply
#62
CornyOC
Chrispy_Rubin is next gen architecture name, so it'll be the same silicon as the 6090 by the sounds of it. Whether it's the actual 6090 or a new model range for prosumer AI use is anyone's guess.

The 3090 and 4090 were definitely marketed by Nvidia as a premium gaming card that was also good for work.
The 5090 was absolutely not marketed as a gaming card. Even though it's obviously that, Nvidia's series of pre-launch and launch events were selling the 5080 as the flagship for gamers and the 5090 as a workstation showcase for all their AI stuff. Remember Nvidia is an AI company now that still happens to do some graphics on the side....
Seems a bit early for an actual RTX card, if 50xx Super get released at least in q2 then maybe. There's no way Nvidia will not launch Super prior, since those are the real GPUs, what we got so far are just detuned GPUs, expect for 5090.
Posted on Reply
#63
modmax
The RTX 5090 uses a chip rated -350 which means it is a chip with disabled shaders. The maximum expression is 400 but you will also find 350, 450 due to RAM modifications or other, which already makes the name inappropriate. 4000 dollars to overclock a "damaged" chip... incomplete I would say that pure speculation... I am nobody the 5080 is the best choice completely active chip...... but that's just my opinion
Posted on Reply
#64
Red Hood
Chrispy_Why not. But it MUST have AI in the name. Google, Microsoft, Meta, AMD, Intel, AMD, and Qualcomm already have products with AI in the model name, so this will need to have at least TWO to make it clear that it's an AI product first and foremost.

Nvidia Titan Rubin AI 8000 Pro (RTX AI Edition)
120GB GDDR7 640 bit

If it's another massive 202 chip then there will probably be a 6090 48GB/64GB 512 bit cause of the low yields, most of which will end up in "ai" workstations and servers
Posted on Reply
#65
KrazyT
Could / Should have been :

ASUS GeForce RTX 5090 ROG MATRIX Priced at $1,000, Only 3,999 Units Available

Posted on Reply
#66
Chrispy_
CornyOCSeems a bit early for an actual RTX card, if 50xx Super get released at least in q2 then maybe. There's no way Nvidia will not launch Super prior, since those are the real GPUs, what we got so far are just detuned GPUs, expect for 5090.
I mean everyone is guessing right now - even Nvidia.

The design looks to be finished, Rubin CPX must already taped out as engineering/validation samples at this point so much of it is already set in stone. For both the 50-series Super models and 60-series Rubin, Nvidia will release when the market conditions make them the most profit (i.e when existing inventory runs dry to create more demand, and when GDDR7 pricing doesn't suck.)

They're not in a rush to beat the competition because nobody really has a an announced competitor to the 5090 or Rubin CPX yet. All they would be doing with a rapid launch is cannibalising sales of their existing production stocks.
Posted on Reply
#67
phints
Where was this 9 months ago? This 5090 is a rip off and old news now.
Posted on Reply
#68
ZeDestructor
CornyOCLC has better thermal conductivity than PTM sheet and is cheaper, but it requires skill and preparation. For a product that is about every ounce of performance, I would expect LC with all components around the die properly sealed off.
I'm no longer sure about LM being better given LTTLinus' results where the PTM was better. Feel free to bring your own matched-cooler+load counterexamples.
CornyOCPTM is singular use and a size large enough for a 5090 die would cost about as much as the most expensive 1g LM tube, which is enough for 2-3 uses.
Even with the higher cost and slight loss of performance, not having to deal with LM, not even with just removing it (which I might have to do to my laptop at some point in the far future), would be more than enough to send me to PTM.
Chrispy_Why not. But it MUST have AI in the name. Google, Microsoft, Meta, AMD, Intel, AMD, and Qualcomm already have products with AI in the model name, so this will need to have at least TWO to make it clear that it's an AI product first and foremost.

Nvidia Titan Rubin AI 8000 Pro (RTX AI Edition)
Call it the tAItan :P
Posted on Reply
#69
vekspec
Liquid Metal again? Should just use PTM like in their other GPUs this generation. Given their history of LM in their RoG laptops…eh.
Posted on Reply
#71
FoulOnWhite
Just watched Der8auers review of this card, pretty good but the liquid metal application was a mess.
Posted on Reply
#72
SkullFox
CornyOCLC has better thermal conductivity than PTM sheet and is cheaper, but it requires skill and preparation. For a product that is about every ounce of performance, I would expect LC with all components around the die properly sealed off.
very marginal gains. not enough to outweigh the risks and complexities. also it needs to be replaced much sooner than PTM.
Posted on Reply
Add your own comment
Nov 21st, 2026 02:13 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts