Monday, September 21st 2026

NVIDIA GeForce RTX 60-Series Reportedly Slips to 2028, Per New Leak

Kopite7Kimi, a leaker with a strong track record on NVIDIA hardware, says the company's next-generation GeForce RTX 60-series based on the Rubin architecture is a 2028 product. "Unfortunately, GR20x is a product from 2028. That means they have postponed again," the leaker posted on X. Their earlier estimate was the second half of 2027, so this is another "delay," if you can even call it that. NVIDIA hasn't said anything official about RTX 60-series timing, so treat it as a rumor for now.

Interestingly, this new leak seems to align with what prominent leaker Kepler_L2 claimed a few days ago: that Rubin gaming GPUs would only arrive in 2028. It also goes against leaker Moore's Law Is Dead, who said about a week ago that an NVIDIA source told them gaming Rubin was greenlit for the first half of 2027. Two leakers against one doesn't settle anything, and we don't know who is right yet, but the 2028 timing does seem more likely given the current industry conditions. Of course, these plans can change before 2028, as the memory supply situation is currently very volatile, and hardware roadmaps are changing almost on a weekly basis.
As for why Rubin seemingly slipped to 2028, the culprit seems to be what we all expected. Memory and TSMC capacity are stretched thin, and data center hardware makes NVIDIA far more money than GeForce does. Apparently, the RTX 50 Super refresh is also stuck over pricey 3 GB GDDR7 modules. All this likely adds up to a long Blackwell cycle, NVIDIA still may manage to squeeze in the Super launch somewhere in 2027, unless it is canceled outright.
Source: Kopite7Kimi on X
Add your own comment

125 Comments on NVIDIA GeForce RTX 60-Series Reportedly Slips to 2028, Per New Leak

#101
Thunderslat
JustBenchingIt's not the same thing though. Have you ever wondered what would be required for Nvidia to stream games to 100 - 200 million simultaneous players? That would take more bandwidth and networking than the entire internet currently has. Let alone the actual GPU / cpu hardware required.

Software had to go sub because if one company did it, every company had to. You cant sell a 3k$ software when your competitor offers it for a 50$ subscription.
The entire Internet? Come on. You do know there is a ton of streaming going on all the time? Gaming uses more bandwidth but give it time it will get there if necessary.

Also the hardware is another benefit. Rather than having to sell a GPU to every gamer, you only need to have enough for every concurrent gamer at any given moment. The rest of your wafers can go to higher margin products.
Posted on Reply
#102
JustBenching
ThunderslatThe entire Internet? Come on. You do know there is a ton of streaming going on all the time? Gaming uses more bandwidth but give it time it will get there if necessary.

Also the hardware is another benefit. Rather than having to sell a GPU to every gamer, you only need to have enough for every concurrent gamer at any given moment. The rest of your wafers can go to higher margin products.
Most streaming platforms that I know off have caps at 5 and 10 mbps, like twitch and Netflix. Gaming needs 10 times that but most importantly, unbuffered. I can't see it happening anytime near
Posted on Reply
#103
Thunderslat
JustBenchingMost streaming platforms that I know off have caps at 5 and 10 mbps, like twitch and Netflix. Gaming needs 10 times that but most importantly, unbuffered. I can't see it happening anytime near
They can do more. Netflix recommends 15Mbps for 4k which I think is a shit bitrate but I guess people are used to it since most don't invest in Blu Ray. With fiber becoming ubiquitous it's not that hard of an ask.
Posted on Reply
#104
JustBenching
ThunderslatThey can do more. Netflix recommends 15Mbps for 4k which I think is a shit bitrate but I guess people are used to it since most don't invest in Blu Ray. With fiber becoming ubiquitous it's not that hard of an ask.
I think you are looking at it from the perspective of the user. Yes speeds of 100mbps are aleasily accessible nowadays, heck I have 300 for 19 euros. But the networking from the streaming platform perspective to offer unbuffered 100mbps streams to 100million+ users is insane. That's literally more data than the entire internet currently
Posted on Reply
#105
jspinks231
Or some I want nvidia to do something alittle more budget friendly they still haven't came down on the 6gb rtx3050 the windforce is probaly nice though..or see how supported the intel gpu is with the arc a380 even better kinda deals i like. As you can see I've needed nothing though and the horsepower edition reference vega56..yeah this one bench's soaring high still, 1080 like one. It's not worth much now your fav game and a four loko probaly maybe pick ya up the sprite and juice one lol.
Posted on Reply
#106
Thunderslat
JustBenchingI think you are looking at it from the perspective of the user. Yes speeds of 100mbps are aleasily accessible nowadays, heck I have 300 for 19 euros. But the networking from the streaming platform perspective to offer unbuffered 100mbps streams to 100million+ users is insane. That's literally more data than the entire internet currently
So a bit over 1/3 of the US will be playing AAA games demanding 100Mbps (what happened to 50?) at the same time will be a thing? Come on. You're wrong, it happens.
Posted on Reply
#107
JustBenching
ThunderslatSo a bit over 1/3 of the US will be playing AAA games demanding 100Mbps (what happened to 50?) at the same time will be a thing? Come on. You're wrong, it happens.
Worldwide? Yes. That's how many concurrent gamers there are
Posted on Reply
#108
Raiden85
So if there’s no 6090 until 2028 that means my 4090 would be the second fastest Nvidia consumer GPU for almost 6 years – that’s just insane 😄
Posted on Reply
#109
mikesg
Nvidia and AMD are testing Intels new 18A node (2nm) which is currently better than TSMCs N2P because Intels has backside power delivery.

Both Intel 18A and TSMC N2P have the new transistor type (Gate-all-around) which is a big jump in performance, but so is backside power delivery (powervia) and TSMC doesn't have that. You can't release a two year product range midway through the maturity of a next-gen node.

AI will spike again in 2028, and consumer GPUs will be the second-hand bad yields of all the AI chips, and still cost $1-3k.
Posted on Reply
#110
Thunderslat
mikesgNvidia and AMD are testing Intels new 18A node (2nm) which is currently better than TSMCs N2P because Intels has backside power delivery.

Both Intel 18A and TSMC N2P have the new transistor type (Gate-all-around) which is a big jump in performance, but so is backside power delivery (powervia) and TSMC doesn't have that. You can't release a two year product range midway through the maturity of a next-gen node.

AI will spike again in 2028, and consumer GPUs will be the second-hand bad yields of all the AI chips, and still cost $1-3k.
Explain to me why Intel is using N2 for their high end NVL then and only using 18A on the low end stuff?
Posted on Reply
#111
TumbleGeorge
ThunderslatExplain to me why Intel is using N2 for their high end NVL then and only using 18A on the low end stuff?
Hmm "low end" stuff work with close and with 6GHz turbo.
Posted on Reply
#112
kazuviking
ThunderslatExplain to me why Intel is using N2 for their high end NVL then and only using 18A on the low end stuff?
bLLC is on 18A tho?
Posted on Reply
#113
Thunderslat
kazuvikingbLLC is on 18A tho?
I don't think so. I should've said the lower core count chips are supposed to be 18A, whereas the higher ones are n2. Last I ready anyway. I think bLLC is limited to higher core count chips like 8 + 16 and 2 * (8 + 16).
Posted on Reply
#114
ToxicTaZ
Nvidia has an Evil monopoly going on.

- Nvidia has 90% market share.
- Nvidia has no competition!
- Nvidia $5 Trillion dollar company!
- Nvidia is years ahead of its compilation.
- 2nd-Gen Transformer Model works on all RTX cards.


AMD best RX 9070XT 16GB slower than the old RTX 4080 Super 16GB (AD103) Technology from 2022.

MSI dropped AMD 9000 Series and AMD dropped backwards compatibility driver support for RDNA3.

RTX 6080 32GB will outperform the RTX 5090 32GB.

RTX 5080 Super 24GB @ces 2027
AMD has some huge shoes to fill.
Posted on Reply
#115
Hecate91
ToxicTaZNvidia has an Evil monopoly going on.
They sure do if you look at the AI bubble which a majority of is funded by Nvidia.
ToxicTaZ- Nvidia has 90% market share.
- Nvidia has no competition!
This isn't a good thing by the way.
ToxicTaZAMD best RX 9070XT 16GB slower than the old RTX 4080 Super 16GB (AD103) Technology from 2022.
So is the 5070Ti, a $1,200+ GPU slower than the 4080 super that launched for $999.
ToxicTaZMSI dropped AMD 9000 Series and AMD dropped backwards compatibility driver support for RDNA3.
MSI Radeon cards were garbage anyway, and RDNA3 still gets the latest drivers the last i checked.
ToxicTaZRTX 6080 32GB will outperform the RTX 5090 32GB.
Citation needed.
Although I'm sure Jensen will do something like the "5070>4090" claim again.
ToxicTaZRTX 5080 Super 24GB @ces 2027
AMD has some huge shoes to fill.
Again, citation needed.
The rumor is no new cards until 2028, if there is anything new it'll be an extremely expensive paper launch.
AMD fills their shoes fine, although a high end card with RDNA5 would be nice.
Posted on Reply
#116
jdrch
As someone who just set up a 5080 PC, I hope this is true lol.
Posted on Reply
#117
JustBenching
ToxicTaZNvidia has an Evil monopoly going on.

- Nvidia has 90% market share.
- Nvidia has no competition!
- Nvidia $5 Trillion dollar company!
- Nvidia is years ahead of its compilation.
- 2nd-Gen Transformer Model works on all RTX cards.


AMD best RX 9070XT 16GB slower than the old RTX 4080 Super 16GB (AD103) Technology from 2022.

MSI dropped AMD 9000 Series and AMD dropped backwards compatibility driver support for RDNA3.

RTX 6080 32GB will outperform the RTX 5090 32GB.

RTX 5080 Super 24GB @ces 2027
AMD has some huge shoes to fill.
I've given up on them completely. Even if they somehow close the performance gap, good luck with them dropping driver support like a hot potato after a couple of years. I'd like to keep my cards for longer, amds plan is to force an upgrade every 2 years, no thanks
Posted on Reply
#118
kazuviking
Hecate91So is the 5070Ti, a $1,200+ GPU slower than the 4080 super that launched for $999.
Why are you comparing current prices with launch prices? Compare launch with launch.
Posted on Reply
#119
chrcoluk
JustBenchingIt's not the same thing though. Have you ever wondered what would be required for Nvidia to stream games to 100 - 200 million simultaneous players? That would take more bandwidth and networking than the entire internet currently has. Let alone the actual GPU / cpu hardware required.

Software had to go sub because if one company did it, every company had to. You cant sell a 3k$ software when your competitor offers it for a 50$ subscription.
I wish those numbers were representative, I send Macrium a query, and the support staff are all sales clerks now trying to convince me to give up my life time license for a annual license that costs more than the entire lifeline license cost, Annual fees tend to be a horrible ratio compared to previous owned.

Nvidia if shifts to rental will make a killing, I am pretty sure already a lot of AI kit is leased.
Posted on Reply
#120
Krit
kazuvikingWhy are you comparing current prices with launch prices? Compare launch with launch.
Launch prices in first weeks are never good. After a month and further it usually starts to make a sense when prices come down. On average nvidia is about ~ 35% more expensive but now it's over.
Posted on Reply
#121
AleXXX666
lol, does somebody CARE?:kookoo::oops::rolleyes::sleep:
Posted on Reply
#122
TumbleGeorge
AleXXX666lol, does somebody CARE?:kookoo::oops::rolleyes::sleep:
Five sheets comments. like we have sambody cares.
Posted on Reply
#123
mb194dc
Goodbye tech progress we hardly knew ye.
Posted on Reply
#124
Ruru
Trepie
So they have time to release a 5090 Ti with a full chip, starts at 8000EUR?
Posted on Reply
#125
MikOnTopGG
RuruSo they have time to release a 5090 Ti with a full chip, starts at 8000EUR?
I don't know. We will see.
Posted on Reply
Add your own comment
Nov 12th, 2026 06:33 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts