News Nvidia GeForce RTX 4060 Ti 16GB Goes on Sale

This is not much of an "entry gaming GPU", but instead a "super budget AI preview GPU" from nVidia.

People with specific workloads that require more VRAM, even with the paltry bandwidth, will want this card, much to their dismay and annoyance I'd say. This includes game devs on a budget even, maybe?

Regards.
 
"Two and a half years later, roughly the same market segment in terms of price has apparently stepped down a notch on resolution. "

And yet you still gave it a glowing review.
Welcome to the new normal.

I honestly dont know what the hell has happened to the media, but all of them are doing the same thing. Just check Ngreedia Unboxed (among others). Its simply insane and worse is that people these days cant even think for themselves, just blindly believe these kind of reviews.


Well, it's their favorite baby, Nvidia.
Same for Ngreedia Unboxed, Jay2Jensen, Ngreedia Foundry, Linus Ngreedia Tech Tips and all the others.
 
This is not much of an "entry gaming GPU", but instead a "super budget AI preview GPU" from nVidia.

People with specific workloads that require more VRAM, even with the paltry bandwidth, will want this card, much to their dismay and annoyance I'd say. This includes game devs on a budget even, maybe?

Regards.
The problem is that a lot of AI workloads need compute as well as bandwidth as well as VRAM capacity. There might be niche markets where 16GB paired with 88 teraflops of FP16 compute and a 128-bit interface are "good enough," but most companies dabbling with AI should just fork over the money for a more capable GPU.

Like, seriously: Who's going to hire an AI researcher at potentially $5,000+ per month, and then saddle them with crappy hardware just to save $500 or even $1,000? If you want to do AI properly, on consumer hardware, get the RTX 4090. If you're a hobbyist just poking around at AI workloads, fine, maybe a $500 4060 Ti 16GB makes some weird sort of sense. But for non-hobby use, this isn't going to suffice.
 
  • Like
Reactions: artk2219
"Two and a half years later, roughly the same market segment in terms of price has apparently stepped down a notch on resolution. "

And yet you still gave it a glowing review.
Ah, yes, because there's nothing better than a C- grade!

"Class, I'd like you all to listen to this amazing paper turned in by Harrison Bergeron. He nailed the middle of the bell curve on this one! I gave it a 70 out of 100. The rest of you with your A's, B's, D's, and F's should be ashamed. Listen to this lack of reading comprehension: 'And yet you still gave it a glowing review.' Isn't that incredibly uninsightful? You can really feel the mediocrity of that statement."
 
Ah, yes, because there's nothing better than a C- grade!

"Class, I'd like you all to listen to this amazing paper turned in by Harrison Bergeron. He nailed the middle of the bell curve on this one! I gave it a 70 out of 100. The rest of you with your A's, B's, D's, and F's should be ashamed. Listen to this lack of reading comprehension: 'And yet you still gave it a glowing review.' Isn't that incredibly uninsightful? You can really feel the mediocrity of that statement."
3 1/2 Stars is not a C-. You know you just got called on your BS and you don't like it.

0 Stars=F
1 Stars=D
2 Stars=C
3 Stars=B
4 Stars=A
5 Stars=A+

The card is 1 star at best.
 
Last edited:
The problem is that a lot of AI workloads need compute as well as bandwidth as well as VRAM capacity. There might be niche markets where 16GB paired with 88 teraflops of FP16 compute and a 128-bit interface are "good enough," but most companies dabbling with AI should just fork over the money for a more capable GPU.

Like, seriously: Who's going to hire an AI researcher at potentially $5,000+ per month, and then saddle them with crappy hardware just to save $500 or even $1,000? If you want to do AI properly, on consumer hardware, get the RTX 4090. If you're a hobbyist just poking around at AI workloads, fine, maybe a $500 4060 Ti 16GB makes some weird sort of sense. But for non-hobby use, this isn't going to suffice.
I said "people" and not "companies". There's a not-so-subtle difference in that.

If someone wants to just dabble in a little of AI work, they now don't need to shell out 4080 money to get 16GB of VRAM. Someone looking to work with CAD (or media) that uses a lot of VRAM, doesn't need to shell out 4080 money for it.

That being said, I don't disagree with the premise that this card is very anemic even for the most basic of AI grunt work, so I'd like to see that tested when possible. The proof is in the pudding as they always say, no? :)

Regards.
 
The grade of C means average which should correspond with 2.5 on a 5 star scale. Jerred is claiming 3.5 stars is a C- Keep in mind that we are talking about a card that is sometimes beaten by its predecessor which should never be the case, ever.
2.5 stars is 50% which = F, 3 stars is a D, 3.5 stars is a C, 4 stars is a B, 4.5 stars is an A, 5 stars = perfect in every way. Unless the writer specifically defines what each star represents, 0 to 5 stars is standardized to the scale of 0% to 100% where each whole number is equivalent to 20 percentage points.
 
16GB of VRAM is good for Editing movies ... many people will buy this card for 8K movies editing. and even 4K movies can benefit from the VRAM increase but not a much as 8K .
 
Last edited:
  • Like
Reactions: artk2219
It's probably a commercial tactic. they purposely fill the gap between the 4060 and 4070, so people can say "just 100$ more and i can get a 4070 so i will get a 4070"
classic upsale by comparaison

that's also why in cinemas you have the small popcorn at 7$, medium at 7.50$ and large at 8$

If they would just sale the large at 8$ people would find it too expensive, but by putting worst deals side by side, the 8$ popcorn suddenly feels like a good deal.
 
Just check Ngreedia Unboxed (among others).
Lol

Shaz12 from OverclockersUK - HUB is an AMD biased reviewer and whenever the 3000 series won in his benchmarks he used to just brush it off.

I love the internet, the red fanbois call them Ngreedia Unboxed, the green fanbois call them Radeon Unboxed and the blue . . . do they even have actual fanbois . . . if so then they'd call them . . . . hmmm something that refers to both of them.

Lets look at what they said about the latest green releases:

4070 ti - "i'm not really blown away by what's on offer here even if we ignore current retail pricing for previous gen hardware and just focus on the suggest retail pricing"
4070 - "not exactly mind blowing stuff afer about 2 and a half years"
4060 ti - "absolutely do not buy this thing for $400 US it's laughably bad at that price"
4060 - "isn't as laughably bad as the 4060 ti but its still pretty bad and as I said it's certainly not worth paying 300 US for"

And lets not forget when Nvidia blacklisted them because even though their 3090 review was positive they didn't shill it the way Nvidia told them too and then published it without those adjustments.

If HuB are Nvidia shills then they don't seem to be doing a good job at it, same as if they are AMD shills 😵
 
What’s sad is the few things I’ve seen on YouTube the numbers aren’t a whole lot better than the 8gb card. In other words, buy a 6800xt and be happy.

It's what I kept telling people, this card was handicapped by it's memory bus long before it's memory size. Like any settings that would actually need 10GB+ of memory would be running at incredibly bad FPS already.
 
  • Like
Reactions: artk2219
It's probably a commercial tactic. they purposely fill the gap between the 4060 and 4070, so people can say "just 100$ more and i can get a 4070 so i will get a 4070"
classic upsale by comparaison

that's also why in cinemas you have the small popcorn at 7$, medium at 7.50$ and large at 8$

If they would just sale the large at 8$ people would find it too expensive, but by putting worst deals side by side, the 8$ popcorn suddenly feels like a good deal.
Relatively speaking of course, absolutely still a bad deal.
 
  • Like
Reactions: artk2219