Saturday, October 10th 2026
NVIDIA Reportedly Ending the RTX 5090, Reserves GB202 Silicon for Workstation GPUs
NVIDIA has reportedly stopped allocating GB202 silicon to the GeForce RTX 5090, according to leaker MEGAsizeGPU (@Zed__Wang) on X. "No more GB202 for 5090, confirmed," the leaker posted, adding in a follow-up that "all future GB202 supplies are RTX Pro exclusive." GB202 is the largest "Blackwell" graphics chip, and the RTX 5090 doesn't even max it out. The same chip powers the 96 GB RTX PRO 6000 with 24,064 CUDA cores, as well as the recently unveiled RTX PRO 5500, which pairs the RTX 5090's 21,760 CUDA cores with 84 GB of memory. Apparently, NVIDIA would rather put every GB202 it gets into those far pricier workstation cards. The company hasn't confirmed any of this as of the time of writing.
A second leaker, hongxing2020, posted "5090 EOL, 5080 24GB ready" alongside an apparent factory notice saying NVIDIA will no longer sell any RTX 5090 model, China-market D variants included. The same notice says an RTX 5080 24 GB will become the top GeForce card and the 16 GB version will be discontinued. That sounds a lot like the RTX 5080 SUPER, which was rumored with 24 GB of GDDR7 on the same GB203 chip, although it isn't clear if the two are the same card. SUPER cards reportedly reached board partners in July, only to be put on hold over GDDR7 pricing. As for the RTX 5090, I doubt NVIDIA would formally call this an end-of-life (EOL) move, since the chips are simply being redirected to a more profitable product line.RTX 5090 buyers were already having a rough time. The card launched in January 2025 at $1,999, and last month European listings went past €5,000 while Walmart sold out at $4,299. Cards built from GB202 chips that partners already have could keep trickling out for a while, though it's reasonable to expect prices to climb further once that stock dries up. Of course, the RTX 5090 was hardly a gaming card for most people by this point, with AI firms reportedly picking up stock in bulk before it ever reached gaming retail. With the RTX 60-series reportedly slipping to 2028, the RTX 5080 24 GB may stay the fastest GeForce card in production for a long time.
Sources:
MEGAsizeGPU on X, hongxing2020 on X, via Videocardz
A second leaker, hongxing2020, posted "5090 EOL, 5080 24GB ready" alongside an apparent factory notice saying NVIDIA will no longer sell any RTX 5090 model, China-market D variants included. The same notice says an RTX 5080 24 GB will become the top GeForce card and the 16 GB version will be discontinued. That sounds a lot like the RTX 5080 SUPER, which was rumored with 24 GB of GDDR7 on the same GB203 chip, although it isn't clear if the two are the same card. SUPER cards reportedly reached board partners in July, only to be put on hold over GDDR7 pricing. As for the RTX 5090, I doubt NVIDIA would formally call this an end-of-life (EOL) move, since the chips are simply being redirected to a more profitable product line.RTX 5090 buyers were already having a rough time. The card launched in January 2025 at $1,999, and last month European listings went past €5,000 while Walmart sold out at $4,299. Cards built from GB202 chips that partners already have could keep trickling out for a while, though it's reasonable to expect prices to climb further once that stock dries up. Of course, the RTX 5090 was hardly a gaming card for most people by this point, with AI firms reportedly picking up stock in bulk before it ever reached gaming retail. With the RTX 60-series reportedly slipping to 2028, the RTX 5080 24 GB may stay the fastest GeForce card in production for a long time.


228 Comments on NVIDIA Reportedly Ending the RTX 5090, Reserves GB202 Silicon for Workstation GPUs
Edit: I'm not losing any sleep over my 1050 Ti not getting any driver updates, either (before anyone jumps on me for looking through red tinted glasses).
The nuance is also about how RDNA2 iGPUs and handheld chips do rely on these and are still sold brand new to this day. And if you think you'll be smart by randomly saying GT 1030 got its support cut because it's Pascal and is still on sale brand new, you'll be surprised to learn it's currently still in the main branch and actively supported, probably because of its sales volume and the fact it may be actively bought by companies; as opposed to the 1050 Ti which is on security/critical updates only.
Another point of view : imagine if the RTX 4070/3080 got its regular driver support cut short right now for whatever reason related to future tech developments. That's what happened to RDNA 2. Maybe they backtracked, but we don't know for sure. I have my bro's 6900 XT PC in my home and lots of games ranging from 2021 to 2026 won't stop crashing even after a clean install and an overkill PSU replacement.
I still don't know why we're talking about this though. The initial topic is about the discontinuation of 5090 sales to the average consumer. Which is a legit move from a company POV since it's now unaffordable to the avg consumer anyways, may as well just redirect it to workstations and servers for 4 times the price. Now companies will get shafted because they're paying 4 times more for a 5090 today compared to last year. There are always losers in such situations, companies are just losing less compared to us since it's either that or $20k RTX Pro Blackwell.
The defense for this decision is an odd one, given it could set a very slippery slope towards the jacket man giving up the dedicated gaming GPU market completely. They're only discontinuing the x90 card, but the next gen may only be x80 or x80ti at the highest end, and nothing to stop them from pulling another "4080 12gb" again with cutting the die size and raising prices.
Also, when you bought a 5700 XT you knew what you were getting. If you still cling on to it after all these years, despite many games running RT by default, that's on you. Same goes for a 1080 Ti. Crying that old hardware doesn't have new stuff is infantile.Let's specify... RDNA 2 didn't get "put in the closet". It's on a secondary driver update branch, which means it still gets updates, but the core driver will only get updated to fix issues... Of which there's mostly none, because it's an old architecture, so all major issues have been dealt with already. There's nothing to fix. And if there is in the future, we'll get an update.No it's not. I have one, so I know.The 1030 as well.I doubt that's a driver issue, as my 6750 XT works just fine.Except that there is no "either". Discontinuing the 5090 means that there will only be $20k RTX Pro cards above the 5080.
It's like saying that Ford is an unreliable brand because the Pinto was recalled for an unsafe fuel tank placement back in the '80s.
Everybody can pick a random example and magnify it to make it look like a systemic issue. Shall we do that with Nvidia's MFG? ;)
Whether someone knows or not whether RDNA 1 had ML or RT is irrelevant. It aged worse, hence a lot of people still using Turing (cause it can play every game as of now) but are not using RDNA 1, cause it can't.Sure. Nvidia cards get access to the same MFG tech that AMD cards do, so it's irrelevant? But have at it. It's a useless feature for older GPUs anyways since it requires a higher fps baseline.
Anyways, I don't think you are being serious suggesting that GPUs still on latest driver, with hardware still able to support and run properly the current games while having the latest and greatest upscaler isn't a textbook definition of aging better, so im out.
A situation similar to the 4080 12GB may arise, we'll be there to correct them, or they won't be there anyways.
Die size doesn't matter, only final performance numbers matter and price will follow accordingly. 4080 was disappointing at $1200 (€1350 with VAT), 4080 Super corrected that a tiny bit by releasing for $1000 (€1120-1200 with VAT), 5070 Ti achieved the same performance for $750 (€880-1000 with VAT) while having the same die size. I don't mind seeing a 80-tier GPU as flagship again, because naming doesn't matter nowadays except for misleading people. But since that performance tier from both Nvidia and AMD is now extending beyond €900 in most EU countries, I don't expect things to become cheaper until this bubble pops.
AMD restructured and implemented a split driver architecture. They continue to support RDNA1, RDNA2, RDNA3 & RDNA4. The Polaris & Vega are under legacy support that receives critical and stability updates. But the remaining are continued support. For example, RDNA 1 & 2 continue to receive day zero game support, game-specific fixes, and bug fixes, including security patches, etc.NVIDIA GPUs actually age worse because their stingy VRAM allocations cause modern games to stutter, while AMD's massive memory pools and open-source features like FSR keep older cards relevant for years.Because someone made a false claim that needed corrected. And even though they continue to believe they are correct, they are not.
Sure, back to topic then.
For example, explain to me how the heck does FSR make amd cards age better when not only nvidia cards can literally use FSR (the exact same version older radeon cards use) - they can even use the latest DLSS on top of that which is better anyways in terms of image quality?
We have a name for this tactic, it's called gish gallop. You are just throwing 50 different false claims and then pat yourself on the back for a job well done.
The tribalism side of things was old 20 years ago, but still so many seem to see every discussion through the debate scoring lens where everyone has to be on a side. It's so strange to me that some have decided my neutrality is a disguise rather than considering I actually believe what I'm saying.
Anyway, your point is moot:
www.techpowerup.com/353919/amd-announces-fsr-4-1-for-radeon-rx-6000-series-apus-and-xbox?amp
The point isn't moot. It's been 16 months since FSR 4 released, and it's only now backported to 2020-2023 RDNA 2 gpus. DLSS was ported day 1 to 2018 nvidia GPUs. RDNA 1 - the competitor to turing - will never get it. It also lacks hardware to play some of the modern AAA games. Clearly nvidia age much better but let's keep pretending otherwise
Edit: Not to mention it's off topic, too.You have FSR 4 now, earlier than AMD promised, so it is moot. Spewing this crap everywhere won't make AMD release the backport at an even earlier time in the past. Also, please keep conveniently ignoring frame generation and MFG that Nvidia can't be asked to backport, even though the hardware to run it is present.
And the thing is, none of this matters here, in a thread specifically about the 5090 potentially getting retired. It's off topic. You acting like the courteous white knight, offering the last word after dropping your "AMD shit" bomb doesn't change the fact.