GeForce RTX 3050 refuses to die as Nvidia plans fifth iteration of its 2022 budget GPU — new Ada Lovelace-powered part suggests the name could even outlive Ampere silicon

(Image credit: Nvidia)

Even some of the best graphics cards come and go, but not the GeForce RTX 3050. The latest version of GPU-Z, a popular information and monitoring utility, shows support for an unannounced GeForce RTX 3050 A graphics card, which appears to be the fifth iteration of the original GeForce RTX 3050.

The initial signs of a GeForce RTX 3050 A appeared last year when it was listed in a PCI ID database as the GeForce RTX 3050 A Laptop GPU. Nvidia later confirmed its existence and specifications, stating that this graphics card was using the newer AD106 (Ada Lovelace) silicon instead of the original GA107 (Ampere) silicon.

The changelog of GPU-Z indicates support for a GeForce RTX 3050 A. There's no word on whether this is a mobile chip, so there's a possibility it could be a desktop graphics card.

There are already four versions of the GeForce RTX 3050 available. The GeForce RTX 3050 4GB (GA107) and 8GB (GA106) launched in 2022, with the 8GB (GA107) model arriving later that year. The now-second-newest version was released last year as the GeForce RTX 3050 6GB. As a result, the GeForce RTX 3050 A would be the fifth variation to wear the GeForce RTX 3050 name.

The AD106 silicon in this new RTX 3050 is the same that Nvidia employs in both desktop and mobile SKUs such as the GeForce RTX 4060 Ti, GeForce RTX 4070 Mobile, and GeForce RTX 4070 Max-Q. AD106 is substantially smaller than the GA107 and GA106 previously used in the GeForce RTX 3050. However, thanks to TSMC's 4N FinFET process, AD106 has a higher density with more room for transistors.

GeForce RTX 3050 A Specifications

Swipe to scroll horizontally

Graphics Card	Architecture	GPU	Die Size	Transistors	Process Type	CUDA Cores	Memory	Memory Interface
GeForce RTX 3050 A	Ada Lovelace	AD106	188 mm²	22.9 billion	TSMC 4N FinFET	?	?	?
GeForce RTX 3050 6GB	Ampere	GA107	200 mm²	8.7 billion	Samsung 8N	2,304	6GB GDDR6	96-bit
GeForce RTX 3050 8GB GA107	Ampere	GA107	200 mm²	8.7 billion	Samsung 8N	2,560	8GB GDDR6	128-bit
GeFore RTX 3050 4GB	Ampere	GA107	200 mm²	8.7 billion	Samsung 8N	2,048	4GB GDDR6	128-bit
GeForce RTX 3050 8GB	Ampere	GA106	276 mm²	12 billion	Samsung 8N	2,560	8GB GDDR6	128-bit

The two 8GB and the 4GB versions of the RTX 3050 have 2,048 and 2,560 CUDA cores, respectively. The 6GB SKU sits in the middle with 2,304 CUDA cores and 6GB of memory. Notably, the Ampere-based graphics card has a narrower 96-bit interface, compared to the 128-bit bus on the other models.

The exact specifications of the GeForce RTX 3050 A remain a mystery. However, the GeForce RTX 3050 A Mobile was limited to 1,792 CUDA cores. This reduction in cores compared to Ampere is understandable, since the Ada Lovelace architecture provides a significant performance upgrade over Ampere. Early benchmarks indicated that the GeForce RTX 3050 A Laptop GPU was very close to matching the performance of the GeForce RTX 3050 Ti Laptop GPU.

Nvidia may have started using AD106 dies in the GeForce RTX 3050 because it ran out of GA106 or GA107 dies and turned to AD106 to fulfill contractual commitments. Alternatively, the company may simply be disposing of leftover or defective AD106 silicon, a strategy that it has often employed previously. AD106 possesses 4,608 CUDA cores, so even two-thirds of the good silicon is more than enough to power a GeForce RTX 3050.

At least Nvidia has made an effort to clearly distinguish the GeForce RTX 3050 A from the other four variations. The release date of the new version remains unknown, however, and it remains to be seen whether a boxed version of this card will ever land on retail shelves.

Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.

See more GPUs News

Zhiye Liu is a news editor and memory reviewer at Tom’s Hardware. Although he loves everything that’s hardware, he has a soft spot for CPUs, GPUs, and RAM.

18 Comments Comment from the forums

Neilbob

Just when the steaming turd shows signs of drying out, someone goes and adds water to it...
Reply
truerock

I needed to upgrade my Nvidia GeForce 690 earlier this year. I figured I'd be waiting for a Nvidia GeForce RTX 5080 I wanted for $1,000 for a long time. So, I purchased a GeForce RTX 3050 6GB for $180 to hold me for a few months. I'm running 4k. 10-bit, 144MHz with it. I kind of feel like I could keep on using it for a long time.
Reply
bit_user

truerock said:
So, I purchased a GeForce RTX 3050 6GB for $180 to hold me for a few months. I'm running 4k. 10-bit, 144MHz with it. I kind of feel like I could keep on using it for a long time.
Oh, just desktop graphics doesn't take much from a GPU.

I have a little AMD Polaris RX 550 from 2017 running a 1440p monitor @ 144 Hz. Even running Google Earth at max quality is nice and smooth on it!

At work, I'm still using a GTX 1050 Ti on a 4k monitor @ 60 Hz. Again, it's more than enough for desktop graphics.
Reply
beyondlogic

Admin said:
GPU-Z, a popular utility for graphics card information, indicates support for an unannounced GeForce RTX 3050 A, apparently the fifth version of the original GeForce RTX 3050.

GeForce RTX 3050 refuses to die as Nvidia plans fifth iteration of its 2022 budget GPU — new Ada Lovelace-powered part suggests the name could even... : Read more

3050 the new 1030 that never dies lol.

amd dont even have anything in the under 200 pound range lol.
Reply
bit_user

beyondlogic said:
3050 the new 1030 that never dies lol.
The bizarre thing is how they're using ADA silicon in a RTX 3000-series part that only ever previously used Ampere. I'm pretty sure they never did that with the GTX 1030.

Why are they doing it?? Are they trying to avoid price erosion of the remaining RTX 4000 cards on the market? Are they somehow trying to capitalize on the RTX 3050's name recognition?
Reply
beyondlogic

bit_user said:
The bizarre thing is how they're using ADA silicon in a RTX 3000-series part that only ever previously used Ampere. I'm pretty sure they never did that with the GTX 1030.

Why are they doing it?? Are they trying to avoid price erosion of the remaining RTX 4000 cards on the market? Are they somehow trying to capitalize on the RTX 3050's name recognition?

i believe these parts could be old laptop chips kind of like what amd did with the rx 6500 xt and 6400

if i was to guess there holding the bottom line with whatever they have on hand. they want to keep the 5000 series above that 200 pound cap they dont want to drop it. basicly youll be fine with last 2 generations if your in that 100-200 cap.
Reply
bit_user

beyondlogic said:
i believe these parts could be old laptop chips kind of like what amd did with the rx 6500 xt and 6400
Those weren't old, just not perhaps originally planned for the desktop. However, they were RDNA2, just like the rest of the RX 6000 lineup. So, not an example of mixing generations, much less re-releasing an existing model number with newer generation silicon!

beyondlogic said:
if i was to guess there holding the bottom line with whatever they have on hand. they want to keep the 5000 series above that 200 pound cap they dont want to drop it. basicly youll be fine with last 2 generations if your in that 100-200 cap.
Okay, but AD106 was RTX 4000 generation. So, it's not eroding RTX 5000 pricing.

I just think this is going to be really confusing, when they want to end driver support for the RTX 3000 series. They're going to need an asterisk on the RTX 3050A.
Reply
Air2004

🎶 We don't die, we multiply🎶
Reply
beyondlogic

bit_user said:
Those weren't old, just not perhaps originally planned for the desktop. However, they were RDNA2, just like the rest of the RX 6000 lineup. So, not an example of mixing generations, much less re-releasing an existing model number with newer generation silicon!

Okay, but AD106 was RTX 4000 generation. So, it's not eroding RTX 5000 pricing.

I just think this is going to be really confusing, when they want to end driver support for the RTX 3000 series. They're going to need an asterisk on the RTX 3050A.

its to inflate the pricing of the 5000 series . if they have to drop the lower end 5000 series to under 200 mark ad106 they may have alot of that silicon left that didnt make the cut.
Reply
JamesJones44

bit_user said:
Why are they doing it?? Are they trying to avoid price erosion of the remaining RTX 4000 cards on the market? Are they somehow trying to capitalize on the RTX 3050's name recognition?
The naming does seem very odd. I understand shifting from GA to AD as it likely all comes down to money, but I'm not sure why they wouldn't just name it 4050 a and just reduce the specs a bit. They could even call it a 4030/4040/etc. since there is nothing being produced for those at the moment but maybe their are plans for those name plates.
Reply

Show more comments