Monday, March 10th 2025
NVIDIA GeForce RTX 5050, RTX 5060, and RTX 5060 Ti Specifications Leak
NVIDIA's upcoming GeForce RTX 50 series mainstream lineup has been leaked, thanks to the well-known leaker kopite7kimi, revealing the complete specification profile of the RTX 5050, RTX 5060, and RTX 5060 Ti graphics cards. The entry-level RTX 5050 features GB207-300-A1 silicon on a PG152-SKU50 board with 2,560 CUDA cores and 8 GB of the older GDDR6 memory across a 128-bit interface at 130 W, while the mid-tier RTX 5060 utilizes GB206-250-A1 silicon on a PG152-SKU25 board housing 3,840 CUDA cores with 8 GB of the latest GDDR7 memory at 150 W. The more powerful RTX 5060 Ti implements GB206-300-A1 silicon with two board versions, PG152-SKU10/15, presumably for two memory configurations. It features 4,608 CUDA cores paired with either 8 GB or 16 GB of GDDR7 memory and a 180 W power envelope.
Most notable is the adoption of next-generation GDDR7 memory technology in both RTX 5060 variants while maintaining the 128-bit memory bus across all three models. While the entry-level RTX 5050 model utilizes GDDR6 memory, the RTX 5060 and RTX 5060 Ti are awarded with the faster GDDR7. The consistent 128-bit interface suggests NVIDIA is leveraging memory compression technologies and GDDR7's increased bandwidth to deliver performance improvements without widening the memory bus. Each tier features a slight increase in CUDA core count, a modest boost over the previous generation RTX 40 series. With the RTX 5050 targeting 1080p gaming, the RTX 5060 is positioned for high-refresh 1080p and entry-level 1440p, and the RTX 5060 Ti is likely aimed at solid 1440p performance across most titles. While we don't know the exact release date, we can expect to hear more about the availability in the coming weeks.
Sources:
kopite7kimi #1, kopite7kimi #2, kopite7kimi #3
Most notable is the adoption of next-generation GDDR7 memory technology in both RTX 5060 variants while maintaining the 128-bit memory bus across all three models. While the entry-level RTX 5050 model utilizes GDDR6 memory, the RTX 5060 and RTX 5060 Ti are awarded with the faster GDDR7. The consistent 128-bit interface suggests NVIDIA is leveraging memory compression technologies and GDDR7's increased bandwidth to deliver performance improvements without widening the memory bus. Each tier features a slight increase in CUDA core count, a modest boost over the previous generation RTX 40 series. With the RTX 5050 targeting 1080p gaming, the RTX 5060 is positioned for high-refresh 1080p and entry-level 1440p, and the RTX 5060 Ti is likely aimed at solid 1440p performance across most titles. While we don't know the exact release date, we can expect to hear more about the availability in the coming weeks.


97 Comments on NVIDIA GeForce RTX 5050, RTX 5060, and RTX 5060 Ti Specifications Leak
Did you care to check if buyers were actual Brits? I wager it's gonna be at least half these cards bought by people from the rest of the world.You're doing it horribly wrong here, sorry. Both AMD and NVIDIA knew tariffs are gonna erect. Both AMD and NVIDIA have EXCELLENT yields. 9070 does ONLY exist to upsell 9070 XT, this is how good things are for AMD. Almost every single one of them has CUs deactivated for tiering reasons, not for factory defects.
NVIDIA don't care (why would they?), AMD did AMD. Navi 48 cards were ready for launch last year when said tariffs weren't remotely a thing. AMD waited till the last possible moment.
Once again: if you plan to reclaim the marketshare you must flood the market with products available for sweet prices. Your MSRP should never be fake. There shouldn't be a situation when more than 75% positions worldwide are sold for 120+ % MSRP. Not to mention how ugly it is to advertise Navi 48 with sweet photos of reference models that doomed to never exist. I hate everything about it.
And btw, we had several times more RTX 3080 GPUs on the launch week than we had 9070 series combined. That still was a paper launch.
//also RT performance in Navi 48 is only impressive if you've never seen anything better than a 2080 Ti
1. If yields are good, then where are the Nvidia cards? Why do we have news of the 5050, 5060 and 5060 Ti releasing on cut-back dies? Where are the full ones?
2. By releasing early, AMD would have given Nvidia the advantage on pricing, and reviewers the chance to bash them for bad drivers. The 9070 XT wouldn't have seemed such a good deal with bad drivers.
I'm also worried about the 5060 8GB - the 4060 8GB has been a problematic sell and already been thrown under the bus by a couple of developers. VRAM isn't expensive, and a 16GB card with low bandwidth isn't a threat to Nvidia's market for AI cards.
For creation, the 5050 would have a warm place in my rack ONLY in HHHL.So even if it needs 6-pin PCI-E, this is it. There's nothing to make guys like me consider the others.
Nvidia needs to work on their product strategy. SKUs all over the place, chips MIA, ROPs DOA, etc.
If yields are so good for AMD and TSMC is at MAX production, these 5050/5060 chips are a scam.
I had an MSRP card order cancelled after checkout was complete just so the retailer could immediately relist it for 15% more; This was between 14:00 GMT when the launch officially happened and 14:03 when my completed order was cancelled and Paypal confirmed that the retailer had cancelled the order at 14:01 GMT, less than 60 seconds after launch.
It wasn't a stock issue, it was a retailer scalper issue. Stock at £699 persisted for another two hours after that but I wasn't interested in a $699 9070XT, I was interested in a $569 9070XT.
I'd be salty about it if I was desperate for a graphics card, but I'm not. I already have a 5070Ti, a 7800XT, and a couple of 4060Ti cards at home so it's not really an issue. I appreciate that many people have skipped a generation or two more than they'd have liked due to scalping and poor pricing, so it must be extra irritating that most of the 9070 stock was scalped, botted, or otherwise unavailable at the promised price immediately upon release.You're in the UK buying a GPU made by a Taiwanese-owned company in China. The US tariffs affect the people in the US importing goods from China, but not Nvidia who export IP to TSMC in Taiwan and then sell from Taiwan to Taiwanese firms like Asus and Sapphire.
The US tariffs shaft people in the US, but that's not a reason to accept scalping of GPUs outside of the US.
as for rtx 5060 = rx 6700 xt + maybe 5%
rtx 5050 = rtx 4060
this is what happens when one corporation has a monopoly on the market
A VRAM-cripple ends-up beeing a FPS-cripple sooner or later. Remember this you 6, 8, 10 and 12GB NVidia customers.
At a point, when a behemoth company no longer cares if a particular segment purchase their products or not, basically grants them the license to do what ever they want. Nvidia is going mad with prices, giving you less for more.
Nvidia’s strategy became known with the 40 series. Make the top product extremely powerful and jack the price up.
And 16gb version at least 500£
Where else you buy GPU?
Theoreticaly you could get 9070 for 529£
But it was by no means cheap.
More like ok price.
Seems to me there is a lot of people buying PCs plus ai Scalpers and be sure monopol will get evry£ they
Also 5060 12Gb is rumored, maybe just super variant or maybe there is no 8Gb 5060 and its 12GB realy
But lets be real here.. u only need to buy GPU once and use it next 2years or longer. and still u can sell it.
its not the most expensive thing in the life,not even close.
Also its 100% optional just like gaming is.
Some ppls just want to get everything very cheap or free... thats also greedy
8GB is paltry. It's simply not enough for anything above 1080p gaming.
What's interesting however is the Steam surveys show mst users either have a xx50 or xx60 GPU - not the higher end models.
I'd guess that's because of budget gaming laptops and desktops.
The switch to the older memory was probably because they're already using so much for the higher model cards; I doubt it was part of their original plans (but I obviously dont know that for sure) but I've heard that NVIDIA already bought out most of the GDDR7 memory modules just for the production of 50 series. Could be part of the many reasons why stock is so terrible. (Don't have a source so take that with a ton of salt)
xx60 being an HRR equivalent of a xx50 1080p/60 maybe almost happened once in the past decade, but only if you bend definitions. And iGPUs have replaced the old 50 class anyway.
Using a very lenient definition of high refresh being 100fps, you're asking this 60 series to be 67% faster than the 50 series. Using 120 fps as HRR, that means twice as fast.
Last few '50 series:
The 3060 is only 34% faster than the 3050
The 1660 is only 49% faster than the 1650
The 1060 3GB is 76% faster than the 1050 (beats the lenient definition)
The 1060 6GB is 61% faster than the 1050 Ti (falls short)
The 960 was all of 12% faster than the 950. (oof)
iGPU's are debatable. Some of the ones Ryzen has are great, but I don't know if I'd go as far to say as they've replaced the 50 class. Then again, the last one I remember being super impressive was one that was about on par with a 1650 (dont know how long ago that was), my knowledge of iGPU's is very limited.
Also to be fair for the other things; I've had 2 shots of bourbon with nothing on the stomach as I was starting to feel it when I typed that out. My bad.
The regular desktop 3050 8GB is 28% faster than the desktop 3050-6GB.
So a desktop 4050 with similar specs would already be about 10% faster than the desktop 3050 8GB, but with much lower power use.
The article's desktop 5050 specs beat the laptop 4050 in memory bandwidth and especially power limit so it's reasonable to expect the 5050 to be 25-30% faster than the 3050.
Then there was the 1030 (2 GB VRAM_ that came out in 2017 followed by the 1630 (4 GB VRAM) in 2022. I was surprised that the 1630 has GDDR6 memory and if there is going to be a 5030 with GDDR6 memory, I'm thinking it will either have 6 GB of VRAM or 8 GB of VRAM.
The 1630 was also a good answer to the Steam Deck given that it had the same number of shader cores but about 100 Mhz faster clock than the Steam Deck. My guess is that the 5030 would have either 640 (like my old GTX 1050) or 768 shader cores (like the 1050 Ti). Speed would be over 2Ghz (should be at least 2.5 GHz) in that case which would get it closer to the PC gaming handhelds (have a Legion Go and the iGPU can hit 2700 Mhz).
The 3050 seems to be similar to the Xbox Series S specs especially the GDDR6, memory speed, and bandwidth. The rumored 5050 specs have GDDR6, same bandwidth as the 3050, and less CUDA cores than the 3050 but the memory speed is closer to the PS5's memory speed.
And back to the 5030, I just discovered that there is a GTX 4010 which has 768 CUDA cores...basically a faster Xbox One gpu. Perhaps the 5030 will have half the cores of the 5050 and that even makes the 5050 look even more garbage. The 5030 in that case would have the same number of shader cores as the Xbox Series S. 5050 has the same number of shader cores as the PS5 and similar memory speed but half the bandwidth at 224 GB/s.
I main in fighting games and Street Fighter 6's recommended specs list the 2070 which has 448 GB/s just like the PS5 (min is the GTX 1060). The 5060 would fit the bill for a next generation minimum spec build but I believe the 5070 would be a better fit for that. My RTX 3060M gets the job done for this generation's fighting games so I have no need for the RTX 50 cards.
I don't think the entire RTX 50 cards are worth buying and I would tell anyone to wait for the RTX 60 cards.