Thursday, March 5th 2026

NVIDIA GeForce RTX 3060 Could Return Mid-March

The NVIDIA GeForce RTX 3060, a mid-range GPU now two generations old, is reportedly returning to NVIDIA's supply this month. According to Chinese Board Channels, NVIDIA is planning a mid-March restock of the "Ampere" GPU, aligning with earlier rumors that suggested a Q1 2026 revival. Interestingly, it is unclear which version of the RTX 3060 will be reintroduced—whether it will be the original 12 GB model with a 192-bit wide memory bus or the newer 8 GB variant with a 128-bit bus. NVIDIA's decision to bring back this older SKU is puzzling, especially considering it is two generations old and comes amid memory supply chain shortages. However, this older SKU uses GDDR6 memory, which might be more readily available as the newer GDDR7 is being used by modern "Blackwell" GPUs and the upcoming "Rubin CPX" accelerators.

Why NVIDIA has chosen the RTX 3060 instead of a newer model like the RTX 4060 remains uncertain. Speculatively, it could be due to the fact that the RTX 4060 is based on the same NVIDIA 4N foundry node at TSMC as the current RTX 5060, while the RTX 3060, along with the rest of the GeForce "Ampere" generation, is built on the Samsung 8N (8 nm DUV) foundry node. Additionally, Board Channels note that GeForce RTX 3060 models from various brands will start arriving soon, which means NVIDIA's add-in card partners are doing much of the heavy lifting to bring back this SKU, with NVIDIA only supplying the GPU die and memory as an installation kit. AICs could start marketing this GPU again or just quietly add it to their websites. We are waiting a few more days to see how the re-launch unfolds and which SKUs we end up getting. Finally, the most important factor for considering this GPU when modern alternatives exist is the pricing, which will dictate its sales.
Sources: Board Channels, via VideoCardz
Add your own comment

38 Comments on NVIDIA GeForce RTX 3060 Could Return Mid-March

#1
outlw
Calling it now: 6 GB on a 96-bit bus :p
Posted on Reply
#2
Frick
Fishfaced Nincompoop
Not bad if cheap enough, but I'm betting it won't be.
Posted on Reply
#3
Vayra86
While I'm fine with Nvidia or AMD utilizing older nodes to make us GPUs, do make us actual new GPUs pls. Its just fine if you bake a 250W x60 on 7nm as long as it performs, but it must be latest gen.

And hey, since Nvidia has AI doing their floorplans for them now, come on, why isn't this a thing yet. After all you've got 30k engineers twiddling thumbs now because the AI is writing drivers already.

:rolleyes:

For the people saying or even thinking this 3060 re release is a good thing... yeah. I know you're desperate, but the answer is obviously no. Just get a second hand 3060 for less then, and force Nvidia to deliver an actual new product if they want to sell something. Surely you're not going to fork over new-GPU-money for something two generations old now, I hope? It falls in the same line as pre ordering: don't be an idiot.
Posted on Reply
#4
dgianstefani
TPU Proofreader
Vayra86While I'm fine with Nvidia or AMD utilizing older nodes to make us GPUs, do make us actual new GPUs pls. Its just fine if you bake a 250W x60 on 7nm as long as it performs, but it must be latest gen.

And hey, since Nvidia has AI doing their floorplans for them now, come on, why isn't this a thing yet. After all you've got 30k engineers twiddling thumbs now because the AI is writing drivers already.

:rolleyes:

For the people saying or even thinking this 3060 re release is a good thing... yeah. I know you're desperate, but the answer is obviously no. Just get a second hand 3060 for less then, and force Nvidia to deliver an actual new product if they want to sell something. Surely you're not going to fork over new-GPU-money for something two generations old now, I hope? It falls in the same line as pre ordering: don't be an idiot.
If it's $150-170 it's not so bad for a new card that supports DLSS. Old 3060s you don't know if they've been heavily used or mined on. $200 or above and it's DOA though. The 5050 8/9 GB can be had for less than $250 if you wait around, so a 3060 isn't attractive unless it's super cheap.
Posted on Reply
#5
Vayra86
dgianstefaniIf it's $150-170 it's not so bad for a new card that supports DLSS. Old 3060s you don't know if they've been heavily used or mined on. $200 or above and it's DOA though. The 5050 8/9 GB can be had for less than $250 if you wait around, so a 3060 isn't attractive unless it's super cheap.
How far does DLSS support go on Ampere nowadays then?
Posted on Reply
#6
dgianstefani
TPU Proofreader
Vayra86How far does DLSS support go on Ampere nowadays then?
Pretty good, I used Transformer on my 3080 Ti for about a year, the new 5070 doesn't feel any different, besides being able to turn on FG in WT, so I can run RT/PT with the sliders maxed out at 3440×1440/240 (base frame rate is around 80-120) but I'm saving ~150 W. I'm interested to try out dynamic FG, so it only turns on when needed, at the level required to hit FPS cap and stay there.

Like I said, a brand new DLSS card for $150-170 is a seriously attractive option, considering the used cards are around $100-150 on ebay.

The main thing isn't just framerates though, turning on any level of DLSS is basically the only competent anti-aliasing, and games look terrible without it. FSR 4 comes close at least.

I don't think I tried 4.5 Transformer on the 3080 Ti, so the likelihood is that has a performance penalty, but should still be better than native in most respects.
Posted on Reply
#7
MarkoB88
Nvidia you are drunk go home
Posted on Reply
#9
_roman_
100€ card, why not
+20€ for fancy cover, etc.
anything else is not reasonable

the new nvidia 750 hopefully.
Posted on Reply
#10
Youseethis
_roman_100€ card, why not
+20€ for fancy cover, etc.
anything else is not reasonable

the new nvidia 750 hopefully.
The 3060 12gb, easily the new 750 ti 2gb.
Posted on Reply
#11
igormp
The 3060 12GB was the BEST budget GPU for ML given its VRAM and was reasonably fast. I still have mine in my 2nd desktop running smaller models.
If it comes at the right price, this will be more than welcome to many folks.
Posted on Reply
#12
NutZInTheHead
I know its not EVGA coming back but I found it funny the picture is of a EVGA card :D
Posted on Reply
#13
Unregistered
Hopefully this means they will back port DLSS and frame gen to the 30xx series.
#14
bonehead123
NutZInTheHeadI know its not EVGA coming back but I found it funny the picture is of a EVGA card :D
Yea, I noticed that too. Perhaps nGreediya discovered a hidden stash of them in a shuttered warehouse or someone's garage somewhere & bought them all up for resale :D

Kinda funny, seeins how I just bought a like-new EVGA 3060/8GB for $110 for a yougin/newbie gamr boi in the family...
Posted on Reply
#15
lexluthermiester
outlw6669Calling it now: 6 GB on a 96-bit bus :p
That's a 3050..
Posted on Reply
#16
mb194dc
Absolutely laughable situation, broken drivers and cards from 5 years ago being re-released! Mark the date, this must be near peak LLM bubble.
Posted on Reply
#17
outlw
lexluthermiesterThat's a 3050..
It might have been a RTX 3050 in 2021, but now is the perfect chance to up-brand it :p

Really though, the RTX 3050 was still 8GB over a 128-bit bus... but with ~71% of the RTX 3060's cores.
There was a RTX 3060 3840SP variant, though, with the full GA106 chip enabled, and only 6GB VRAM.
Perhaps this will be the variant they bring back... though they will probably go with a lower enabled core count, so that they can use all their new dies 🤔
Posted on Reply
#18
neatfeatguy
Oh snap! Let EVGA know they can get back in the GPU business for those needing the nostalgia of a decent GPU manufacturer.

I still run my EVGA RTX 3060, it doesn't see any gaming on it anymore, but it still going in the server computer.
Posted on Reply
#19
vbq7qK68eyYAH4iR
Mere Mortal
I thought it was memory allocation, not silicon/TSMC that is the bottleneck. Is supply of GDDR6 really that much less constrained that this makes sense?

On the upside, if this does happen and new products are getting release, Ampere will be getting driver upgrade until 2030 or beyond.
Posted on Reply
#20
Panther_Seraphin
vbq7qK68eyYAH4iRI thought it was memory allocation, not silicon/TSMC that is the bottleneck. Is supply of GDDR6 really that much less constrained that this makes sense?

On the upside, if this does happen and new products are getting release, Ampere will be getting driver upgrade until 2030 or beyond.
There is constraints on both sides. Memory is the obvious one, but the GPU is constrained because of the demand for data center class skus such as the MI350 or B200 parts.

Nvidia/AMD would much rather used their wafer allocations for those higher margin parts than consumer grade GPUs.

But older nodes don't have that demand from data centers but may be "ok" for the mass consumer markets. Just think about how many people are playing on midrange Vs the titans/xx90s
Posted on Reply
#21
dgianstefani
TPU Proofreader
Panther_SeraphinThere is constraints on both sides. Memory is the obvious one, but the GPU is constrained because of the demand for data center class skus such as the MI350 or B200 parts.

Nvidia/AMD would much rather used their wafer allocations for those higher margin parts than consumer grade GPUs.

But older nodes don't have that demand from data centers but may be "ok" for the mass consumer markets. Just think about how many people are playing on midrange Vs the titans/xx90s
Yeah for anyone on a 1060/580 a 3060 is a big upgrade.
Posted on Reply
#22
Chrispy_
There's no way this will be the 12GB version, so it's the bandwidth-starved 8GB version that's barely faster than a 3050.

I'm guessing this is because Samsung have affordable capacity going spare to make Ampere chips, while TSMC are basically sold out for anything Ada or Blackwell . If it was solely a GDDR7 pricing/availability issue then we'd be getting more 4060 or 5050 cards since they use GDDR6 still.
Posted on Reply
#23
N/A
Makes no sense. That card is slow, misses the mark. Best produce a 3070 ti with 192 bit and 12 GB.
Posted on Reply
#24
Chrispy_
N/AMakes no sense. That card is slow, misses the mark. Best produce a 3070 ti with 192 bit and 12 GB.
GDDR6X isn't in production any more, there's no supply of it so you won't get higher end cards that need GDDR6X.

In saying that, the 3070 used regular GDDR6 at first, so they could make it but I'm going to guess that they won't because the 3070 performs like a 5060 and Nvidia have no problem meeting demand with the 5060, and that's plenty profitable with it's paltry little die, simple little 128-bit PCB, miniscule VRMs and cooler, and only half the PCIe lanes to worry about.

No, I'm willing to bet the farm that Nvidia is just trying to make a low-end card they can sell for $250 street price, since the 5060's street price has climbed to $350-400 now. There's a demographic of people with $250 to buy a graphics card and Nvidia doesn't want them buying used GPUs or competitor GPUs like the Arc B570, so they need to sell something at that price point because a good fraction of buyers are too ignorant to do anything except buy a new Nvidia card that fits their budget.
Posted on Reply
#25
lexluthermiester
Chrispy_There's no way this will be the 12GB version, so it's the bandwidth-starved 8GB version that's barely faster than a 3050.
You need to go read some benchmarks. Also, 9GB would be on a 192bit bus. The 1.5GB and 3GB memory chips are 64bits each. Just throwing it out there..
Posted on Reply
Add your own comment
Nov 11th, 2026 03:59 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts