Why does everyone hate the 4080?

[Dirt Chip]

But.. It doesn't. Because the 1,599$ MSRP of the 4090 is just an imaginary number at this point.
First of all, many 3rd party companies sold it for more than that.
Second, the actual current price of a new 4090 is 2,080$ - unless I am gravely mistaken and someone out there does sell it for less. Would appreciate a link.
No one is actually selling them for that price, so whether the MSRP was 1$ or 50,000$ makes no difference right here and right now.
As it stands, the 4090 is 811$ more than I had paid for the 4080.


Cyberpunk is honestly the most demanding one on the list, and even that is sitting at a nice 56 FPS.
I won't be a complete graphics slot (get it?) and demand full max ultra settings. I'm definitely OK with lowering it a bit.
But still, 56 FPS isn't all that bad in my opinion.
Everything else is higher than 60 FPS so.. What do you mean by that sentence?


AMD and Intel have been doing the same thing for years. That's not necessarily bad, it's similar to a tick-tock model.
Another full disclosure, I'm a past Intel employee and have been responsible for verifying the programs which end up verifying the CPUs. So to say I'm not fully invested may have been a misstatement.
I definitely understand the fine details, I just haven't been following them when it comes to GPUs.
Thanks for sharing that detail though, was interesting to know


First of all - thank you for the lengthy comment, really appreciate it!

This intrigues me, as I've heard those claims around, but can't really understand where it's coming from.
According to Techpowerup's testing, the 4090 is a mere 25% faster than the 4080 - and that's only at 4K.
At 1080p & 1440p it is only ~5-11% faster than a 4080.
Why do people consider this to be such an enormous gap in performance? I mean - the price difference is 33%.

View attachment 270287


I don't see it as a scheme or a manipulation of the mind, not at all. The fact is I can't get a 4090 (or even a 4080 now) at MSRP even if I wanted to.
They may be providing very few units to sellers, causing the market to dry very quickly, but that doesn't change the fact the market is dry and if I need a GPU I can't get one.
It being "desirable" doesn't matter much to me, I was just making a point saying they are gone.
Whether it's an elaborate scam or just them being greedy doesn't matter much to an end user. I either have a GPU to play on or I don't.


nVidia is a much more well-known and tursted company and they are dominating the market. Not by a bit, but by a lot.
When a new player attempts to enter the market, they make it worth while by having lower prices on things that match the performance of the competitor.
AMD has definitely been in the market for a while, but mostly at the low-medium end of GPUs, making nVidia the king of high-end and supporting the "Duopoly" you had mentioned.

I'm not saying AMD is the "nice guy" here, but they will definitely sell at better prices just to have more control over the market. it's just how it is.
And when I meant I want to see AMD as a competitor I meant GPUs that can humiliate the 4090 / 4090 Ti to come. I know AMD does well mid-range.

Yes, they'll still price them at a profitable (and even greedy) prices, but not as much as nVidia, if they can do it.

This will either force nVidia to lower their prices, or AMD to lower theirs even further if sells won't be at their desired rate.
Something will have to happen once AMD puts out an actually good GPU.
They had a TON of issues in the past, causing people like me to shun away from them, but I am definitely keeping an open mind here and willing to try an AMD card if their performace:cost ratio is better.

View attachment 270288


I'll ask you this, then.
What does "more for money" mean for you?
If it has triple the amount of VRAM / CUDA cores but the end result is lower FPS, did you get more for your money?
Just trying to understand where exactly is the issue when looking at FPS when selecitng a GPU.
I'd appreciate learning what am I missing and why things like die size, core count, etc.. matter that much over end results (FPS).

Fellings play a major play in this game.
It is a very much emotional decision that for many reflect highly about their own self.

Plus people tend to stick to tier naming more than the company who make the product and decided the name.

Big companies exploit those heumen quality to the max by the help of PR department. People anger and hate when feel exploit.

Now that anger found itself a target, a symbol, in the form of 4080 16GB. By all standard 4080 price too much but from the name xx8x you expect to hold some value vs xx9x. This time there is no value- 25% pref uplifted for 25% more $. Zero value.

Add to that the 4080 12GB incident, another clear story of zero value as above and understand why the rage is on. It is about all the big capitalism companies business practice of exployment, not about 4080.

 

[INSTG8R]

Steve from Gamers Nexus says it best in his review. MEH!

Performance, sure, value? Not so much…

Fellings play a major play in this game.
It is a very much emotional decision that for many reflect highly about their own self.

Plus people tend to stick to tier naming more than the company who make the product and decided the name.

Big companies exploit those heumen quality to the max by the help of PR department. People anger and hate when feel exploit.

Now that anger found itself a target, a symbol, in the form of 4080 16GB. By all standard 4080 price too much but from the name xx8x you expect to hold some value vs xx9x. This time there is now value- 25% pref uplifted for 25% more. Zero value.

Add to that the 4080 12GB incident, another clear story of zero value as above and understand why the rage is on.

Well said!

 

[wolf]

Performance, sure, value? Not so much…

Even Digital Foundry (who I hear constant criticism for being super mega Nvidia shills) multiples times over the course of the review video say that the pricing isn't right, and go in depth into why it's priced badly, I was actually shocked just how many times Richard said it again and again, good. I'm not sure what more OP wants here, the press agree, the users in the know agree, but there's this constant doubling down on why it's priced right. I can't really be bothered writing an essay about it.

 

[AsRock]

Steve from Gamers Nexus says it best in his review. MEH!

Sounded like he called it Meh ( due to pricing ), don't stop it from being a good performer.

 

[evernessince]

This intrigues me, as I've heard those claims around, but can't really understand where it's coming from.
According to Techpowerup's testing, the 4090 is a mere 25% faster than the 4080 - and that's only at 4K.
At 1080p & 1440p it is only ~5-11% faster than a 4080.
Why do people consider this to be such an enormous gap in performance? I mean - the price difference is 33%.

The best example of the raw performance of the 4090 might be VR, where you see a 70%+ boost in performance over the 3090 Ti: https://babeltechreviews.com/rtx-4090-performance-45-games-vr-pro-apps-benchmarked/

TPU's 4K performance number was bottlenecked by their CPU choice. Heck even if they had used a stronger CPU, no CPU on the market can run the 4090 full speed at 4K in many games. This is in part due to it's performance and in part due to the additional driver overhead noted in TPU's review. In non-CPU bound scenarios, the 4090 should be between 30 - 46% faster than the 4080.

You can't be using 1080p and 1440p as examples when comparing absolute performance of these cards as in many games performance is CPU limited, not GPU limited. As CPUs become more powerful, the gap between the 4090 and 4080 will likely increase.

Realistically Nvidia priced the 4080 and 4090 to provide similar performance per dollar and that's not a good thing given most folks don't like flagship pricing. The logic up until now has been that the crazy flagship pricing was fine so long as it didn't affect the rest of the GPU stack. Of course that's not the only reason as I explained above.

nVidia is a much more well-known and tursted company and they are dominating the market. Not by a bit, but by a lot.
When a new player attempts to enter the market, they make it worth while by having lower prices on things that match the performance of the competitor.
AMD has definitely been in the market for a while, but mostly at the low-medium end of GPUs, making nVidia the king of high-end and supporting the "Duopoly" you had mentioned.

I'm not saying AMD is the "nice guy" here, but they will definitely sell at better prices just to have more control over the market. it's just how it is.
And when I meant I want to see AMD as a competitor I meant GPUs that can humiliate the 4090 / 4090 Ti to come. I know AMD does well mid-range.

Yes, they'll still price them at a profitable (and even greedy) prices, but not as much as nVidia, if they can do it.

This will either force nVidia to lower their prices, or AMD to lower theirs even further if sells won't be at their desired rate.
Something will have to happen once AMD puts out an actually good GPU.
They had a TON of issues in the past, causing people like me to shun away from them, but I am definitely keeping an open mind here and willing to try an AMD card if their performace:cost ratio is better.

No, AMD has been making high-end GPUs for awhile. You are gravely mistaken in saying they mostly dealt in low to mid range cards. Go lookup AMD's (and formally ATI) older cards and the corresponding reviews.

You said AMD is not competitive but benchmarks clearly show that they are. Whether Nvidia is dominating or "trusted" is subjective and not really a reply to that. Ask Nvidia shareholders, 970 owners, or AIBs that have to put up with Nvidia's shenanigans including things like the GeForce Partner Program that Nvidia had to pull away from due to blowback if they are "trusted". Nvidia's entire strategy since the 2000 series has been to exploit their customers and partners for maximum wealth, why in the world would you trust them? Every company that has ever worked with Nvidia has come back with a negative impression, just ask Linus Torvalds.

If the AMD issues in the past dissuade you then why are the space invader artifacting with the 2000 series, cards being bricked by a driver bug brought forward via New World with the 3000 series, and now the current kerfuffle with the 4090 and 4080 (given it has the exact same connector) not an issue? It's fine to point out prior issues with AMD but should I also bring a list of just recent issues with Nvidia? Heck it took Nvidia 6 months to fix VR stuttering just recently. I have cards from both brands and people tend to overblow issues from both sides. I almost never have an issue with either. What I do see is a lot of people on the Nvidia side very quick to forget the issues Nvidia does have. It's people defending their purchase and the larger the price tag the more people feel the need to do so. This has been proven in multiple studies.

AMD or Nvidia haven't really competed on price since the 1000 series. Nowadays they price around each other, not against each other. Why would they get into a price war when they can both maintain margins? You can already gain marketshare by offering slightly better value or a slightly better product. There's no need for offering customer outstanding value, especially when they'll then have to top that outstanding value next generation. Better to give customers as little as possible at as high a cost as possible. That's the perspective AMD and Nvidia are looking from. The only thing that might change that is overarching market conditions that force them to lower prices. Hopefully a recession is just that.

 

[Calenhad]

I agree with most other posters here, that the 4080 is overpriced compared to what it offers. Plenty of good reasons have been posted above.

But I need to address a couple bonkers claims in this thread:

If you've followed GPUs for long, including those fine details, you'd know that Nvidia is selling a 103 die, which is historically a xx70 series card, and charging $1200 dollars for it.

Some of us into the fine details still miss the full 100 series dies, which us lowely enthusiasts can't feasibly get our hands on anymore.

You have conveniently skipped the fact that: RTX2080 was a *04 die, GTX1080 was a *04 die, and GTX980 was a *04 die. You have to go back to the GTX780 to find a top tier die in a *80 product, which was released back in 2013. And that die was numbered 110, because Nvidia was always good with numbers.

In fact, Nvidia have never had a *03 die before (I stopped counting at the 200-series).

To further complicate things neither 20- or 30-series had any product using a *00 die. They both started at *02's.

So, just going by die numbers is a flawed argument.

You are paying $3.166 USD per mm2 with the 4080

You are paying $1.275 USD per mm2 with the 3070

WHAT? Repeat after me: Size does not matter! This is like saying 100 grams of apples should cost the same as 100 grams of steak, because they share the same weight.

If you want to count dollar per something physical, at least pick something meaningful. Such as transistor count.

 

[W1zzard]

I've decide to cancel one of the orders of the 4080 and wait for the 7900 XTX. If it disappoints, I'll just wait another damn year and see what'll happen to GPU prices. Sigh.
I'll discuss this with the person I had bought the 2nd GPU for and try to convince them to give it up and wait for the 7900 XTX as well.

dont do that .. if you can afford a product, enjoy it, don't listen to others. it's your money, you worked for it, so you can spend it in any way you want

 

[defaultluser]

its not so bad in-context because it uses a much smaller die size than 4090, its still price withi 40 percent of the big boy ;in the past, cards with such a cut have typicality been a 2: 1 RATIO

1080 ti (102) to 1080 (104) yields a 35 percent performance cut for 60 percent lower price (
2 to 1 ratio in performance per dollar

even though Ada uses one more die,the one used by 4080 is the exact same configuration as p104 used by 1080 yet you still get 2:1 PERFORMANCE/DOLLAR RATUI

 

[qubit]

I remember the good old days when one could buy the next gen product of the same tier* for the same price as the old product and reap the benefits of the performance improvement. At the same time, the old product was heavily discounted until it sold out.

*eg 3080 v 3090 where the new model was the same price as the old model and the old one discounted. I'm just making up the model numbers here to illustrate the point.

 

[Vayra86]

I think its really simple. Patience is running thin and this x80 is yet another card that won't hit any kind of volume markets. Its just priced out of it. They can drop price, sure. They can slot in more cards under and above it, sure. But the 4080 today is what we have, and it does not move the market anywhere. Its complete stagnation, in fact its worse, its making the overall offering worse, effectively lowering perf/$. That is unusual, and a break from the norm.

I'm still on Pascal, and honestly, there hasn't been much progress in perf/$ since. Only regression. The only way you can defend there has been any meaningful progress (making an upgrade... well... not obscene) is if you consider your RT frames valuable. But that, too, is regression in perf/$. IMHO that regression isn't worth what's gained in image quality, not even by a long shot. I have 1,5K on the shelf here to waste on computer parts. I don't jump on this, and I never will, and that's with a 1080 that is getting long in the tooth at 3440x1440. I'm still much happier dialing down settings and accepting lower FPS, than wasting cash on current offerings. That's the market doing its job then... but its still annoying.

And why are people outraged and shitting all over Nvidia? Because team green has absolutely no limit on the greedy practices; sure, companies are in it to make money. But excess is excess, and this is excessive. If these margins are combined with fantastic products priced competitively, everyone is winning. But they're not.

One might start to think Nvidia is actively destroying a market, and desires a move to a different one. Enterprise/datacenter. Even if they aren't, it sure feels that way.

 

[The King]

Sounded like he called it Meh ( due to pricing ), don't stop it from being a good performer.

Does not matter how it performs if its not priced right. The title ask why the hate? Its because of the pricing. So that answers the OP question.

Steve explains in depth why when he gave the example of the CPU market and how crazy fast CPUs have gotten but not priced like crap.
This point was made very clear in the video.

Pricing at the end of the day plays an "important" part of any consumer product value and this offers very poor value thus the hate from reviewers and consumers.

 

[Dirt Chip]

I think its really simple. Patience is running thin and this x80 is yet another card that won't hit any kind of volume markets. Its just priced out of it. They can drop price, sure. They can slot in more cards under and above it, sure. But the 4080 today is what we have, and it does not move the market anywhere. Its complete stagnation, in fact its worse, its making the overall offering worse, effectively lowering perf/$. That is unusual, and a break from the norm.

Yep. It seems that NV 'boiled the frog' too fast this time. It already backlash with 4080 12GB, hopefully the same will happen with 4080 16GB.
I give it a low chance but one can still freally dream.

 

[Arkz]

dont do that .. if you can afford a product, enjoy it, don't listen to others. it's your money, you worked for it, so you can spend it in any way you want

They could, and then regret it afterwards. I'd personally wait for AMDs cards to come out before making a decision.

 

[TissueBox]

I think the chief problem is that value has historically and consistently went up for most generations of the xx80 cards with the exception of Turing/RTX 2xxx.

For example, the 3080 was 66.7% faster than the 2080 at 4k and was $100 less.
The 1080 was 69.4% faster than the 980 at 4k and was 9%-27% more expensive (AIBs $599, FE $699).
The 980 was 31.6% faster than the 780 at 4k and was $100 less.
The 780 was 26.6% faster than the 680 at 1600p and was 30% more expensive. (This was from the same architecture when AMD didn't have a response for the entire 6xx generation, 680 was GK104 and 780 was GK110)
The 680 was 29.8% faster than the 580 at 1600p and was the same price.
The 580 was 16.3% faster than the 480 at 1600p and was the same price.

For most generations, value has gone up substantially because performance went up while price went up less or remained the same. Having a 1:1 ratio for price increase % and performance increase % isn't really something that me, or a lot of people, get excited for. This situation reminds me of the 6xx and 7xx generation because the competition wasn't there so Nvidia could charge whatever they wanted. Thankfully AMD is coming around sooner rather than later this time.

 

[INSTG8R]

They could, and then regret it afterwards. I'd personally wait for AMDs cards to come out before making a decision.

This! because we know the pricing is better all round. Just have to wait to see how it competes with it.

 

[c2DDragon]

Price & Power Consumption.

Half the people on Earth say to use less energy to save the planet, the other half build power angry shits with insane price, there is even a part of those who have fun using tons of fuel launching rockets in space for their own profit.

I don't say I care this much for the planet that I use a bicycle and all but I'm quite sure efforts can be done to avoid having computer components using so much power Oo What will it be in 5 years ? 1200W for the graphic card only ?

 

[Magic Cube]

My opinion about regarding RTX 4080 is that it is just too expensive. Yes performance is good, but still no thank you. Other thing I don't like and didn't like in earlier 3000-series is the power consumption->heat these cards produce. Also I haven't seen really interesting/useful use of raytracing yet. When cheaper cards can run it atleast 60FPS in every game. Maybe then I am interested for now it is just meh.

I am still using GTX 1080, but I don't play games like I did years ago. So these cards are not targeted to me.

In future: 7-Slot gpu with 1200W power consumption with it's own built in/or external power supply.

 

[Vayra86]

And this is why you're not upset about the price, and that's fine. Some of us are fully invested into the fine details, and we know what die is being sold, and care about the memory bandwidth. Some of us aren't just consumers, we're enthusiasts.

If you've followed GPUs for long, including those fine details, you'd know that Nvidia is selling a 103 die, which is historically a xx70 series card, and charging $1200 dollars for it.

Some of us into the fine details still miss the full 100 series dies, which us lowely enthusiasts can't feasibly get our hands on anymore.

The fine details are not what you're saying, unfortunately

History lesson from Kepler onwards (they had a completely different SKU structure <fermi):

- Nvidia was selling GK104 at the top end with Kepler, the 670 and 680; dual GPU GK104 even with the $999,- GTX 690.
- Kepler Refresh/pre-refresh brought Titan as the 100 chip. Stiff competition from AMD pushed them to release 780ti and 780, cut down Titans (GK110).
- Maxwell sold GM204 as a 970 and 980, a highly competitive chip with a famous 500MB of slow memory on the poor man's version. And GM200-310 appeared as a cut down Titan in the Geforce family (about 10% less shaders and still 6GB, exceeding the norm in the market on every aspect with great clocking, making it a long lasting chip).
- Pascal sold GP104's as an x70 and x80 and GP102 (first 102, normalization of a cut down 'pro' chip in Geforce stack) for the 1080ti. Again, the strongest pre-RT offerings and a continuation of what was great about Maxwell, albeit with a somewhat higher price point. Still, the price increase was easily offset by the massive perf jump and perf/w gain. Also, boost was 'perfected' here, and punched way above its weight.
- Turing was the same stack order but additional RT pushed several other metrics to stagnation: perf/w, perf/$ most notably. Cards saw their first TDP increase in the stack since many years. They also saw a double time delay from the past gen in release date. New norms were set.
And then the 'trouble' starts.
- Ampere: uses GA104 had to be relegated to an x60 (!!!) ~ x70ti tier. GA106 appears from x60 (yeah... lol, its a mess) down to 3050, while there are also 3050's with an even smaller GA107. Lots of crossovers like that in the stack, making it clear Samsung yields were absolute shite and the node not great. Power consumption jumped. Cards got bigger. GA102 presented a handicapped set of products with very low VRAM, and had to be used in the 3080 already, instead of just historical x80ti 'top tier' single GPU, and only with more mature node capacity did they produce higher VRAM versions throughout the stack.
- Ada: uses AD102 for x90 only; uses AD103 for x80 16GB only; uses AD104 for a whole range of lower tiers starting at x70. This is the second gen where Nvidia needs to stretch further in their SKU count, not adding stuff on the bottom end (106's, 107's etc.), but rather on the top end. And that is on top of a very good node shrink. This is the first time in GPU history, if I recall, where such a thing has happened; its inevitably going to cause more strain on margin, but Nvidia prices so obscenely, that they can still improve margins on this. That is, if they actually sell.

Seeing the facts like this, its easy to see Nvidia has been diving into a hole since RT was implemented, and needs more spacing/more SKUs in the stack to keep it afloat. More SKUs increases risk and makes the stack more vulnerable to competition, because they are tied to different chips for different products, leading to inefficiency or some chips just getting pushed out of the competitive range. We already saw a version of the 4080 get axed because of exactly that squeeze. With the above history we can understand why that happened: they're supposed to run an x80 card on their 104.

Will Nvidia survive this in the end? Nobody can tell; but its clear the stack order 'escalates' and they need tons of measures and adjustments to keep it going, and we're getting less happy with the results.

 

[Chomiq]

dont do that .. if you can afford a product, enjoy it, don't listen to others. it's your money, you worked for it, so you can spend it in any way you want

I believe this type of logic is what actually got us into the "4080 is $1200" territory.

 

[Dirt Chip]

Fellings play a major play in this game.
It is a very much emotional decision that for many reflect highly about their own self.

Plus people tend to stick to tier naming more than the company who make the product and decided the name.

Big companies exploit those heumen quality to the max by the help of PR department. People anger and hate when feel exploit.

Now that anger found itself a target, a symbol, in the form of 4080 16GB. By all standard 4080 price too much but from the name xx8x you expect to hold some value vs xx9x. This time there is no value- 25% pref uplifted for 25% more $. Zero value.

Add to that the 4080 12GB incident, another clear story of zero value as above and understand why the rage is on. It is about all the big capitalism companies business practice of exployment, not about 4080.

I want to copy past my post from other thred that I find very connecte:

I found a different way out of the upgrade limbo- I waited long enough now that I don't even bother with modern AAA and really don't care about that rat race. I suggest everyone to try- it is very liberating. As NV did, push the un-buy bottom, stick to it for 5 years or so and the magic will begin. And I haven't giving up on gaming at all, not for a day.
I keep my good old faithful 970GTX and play whatever free game EPIC, STEAM or so send me. Many good "old" but new to me titles are still keeping me busy very happily.
You can say that NV\AMD business practice left me no choice or shoved me in that direction and maybe it is true, but in the end they don't see my money and that what's count.
I don't see any productive way to make a change except make the call not to upgrade or to propone it as much as you can. And that practice is DOA because most people don't have the patience to wait- gaming is filling a very important need to them.
I know that and I know that trying to change that very human nature is to battle windmills. It is their escapism and aiming at a long range goal, while sacrificing today, is not a possibility.
Unless united under a very strong symbol, and no such symbol is around, the change will not come from forums like this or from YouTube channels.
Only fierce competition will help.

By un-linking your-self from the hardware you run- that is not feeling crippled by a 'cut-down' GPU, obsolete by having 'outdated motherboard IO', self giving up by compromising on eye candy setting, being fulld by having 'fack e-cores' and most importantly not playing the red vs blue vs green love affair game- I can quite freally choose just what is right for me in the time that's right for me.

A fun fact: Just yesterday I ordered my next system that I've been working on for the past 2 years, mainly by waiting for the right parts for me to become. A nice video editing machine based on 13900k and 64GB DDR5 that will replace my current "dirt sheep" i5-2400 with 16GB DDR3 (I managed to skipped ddr4 and pcie4 altogether, yey). The 970GTX will stay for the time being, until a suitable canidae will show itself, hopefully in 2023 but I'm in no hurry.


And now I will add:
Know this:
If you buy into buliving, trusting and counting on a global company that is listed on a stock exchange in capitalist orianted market- than you are about to get disappointment, frustration, anger and hate sooner or later.
Any attempt to "gain" their trust or favor by "supporting" their product in doomed to failure.
Dont be loyal to a brand, product or tier structure from those companies. It is fundamentally different from supporting a small to medium local company. The language those big tech company are speaking is different from the language common consumer talks.
What many express with the 4080 16GB price is a resoult of counting on the past (getting better value with lower tier products) as a promise for tomorrow. Don't bulive those promises (if exist at all, maybe it is only the person imagination making those promises up) these companies. They own you nothing.
In a deeper sens, people are feeling that they are being denied on a $$$ basis from the thing that they think they are entitled to and the one that deny it from them it is the tech company they are so loyal to. Like getting dumped by the person you love.
By not being able to purchase the new "I-am-entitled-to-it" stuff one might feel he is a lesser person in some way.

So my suggestion to sum it up to whoever cares:
1- Don't be friend with any global company that is listed on a stock exchange in capitalist oriented market.
2- Delay as much as you can any hardware purchase and when you buy- buy purely based on your spacific need at that time without any brand loyalty considerations or future promises.
3- Understand the the hardware in your computer say nothing about you personally.
4- Accept that the current value-is-getting-worse-gen-to-gen situation will continue and will be even worse from year to year. Adapt your buying habits accordingly. The sooner the better.
5- Make love in your heart, Not war.

 

[wolf]

I'm still on Pascal, and honestly, there hasn't been much progress in perf/$ since. Only regression.

I went from a 1080 to a 3080, $1139 AUD at launch for the 1080, and $1399 AUD at launch for the 3080. The problem was buying the 3080 at msrp. Looks like nvidia doesn't want to make that 'mistake' again, so hopefully people voting with their wallets sends a message, imo it clearly already is.

sure, companies are in it to make money. But excess is excess

I dread to see how AMD would price in the top spot lol, but over the years we've seen tastes of it, the benefit of playing second fiddle is doing slightly less greedy things and copping bugger all of the heat for it.

In any case, my gut feel is that AMD'S launch will severely impact 4080 sales, the 4070ti(?) will have to have a drop to $800 USD at a minimum, and hopefully the 4080 to $1000 at a minimum, or replaced with a different SKU to try and save face lol.

What do we think out of these two hypotheticals?

• 4080Ti is a maxed out AD103
• 4080Ti is a more cut down AD102, and the 4090Ti launches as a maxed out AD102

 

[P4-630]

It's the price...

 

[Lei]

Because 4090 blows it away. And is just 30% more expensive
@Omga4000

 

[Keullo-e]

Way overpriced when thinking about its segment. Cut that price to half and it's a good product.

 

[BSim500]

I've been seeing everyone BASHING on the 4080, honestly, for not apparent reason. The main argument is: The 3080's MSRP is 699$, the 4080 is 1199$.

Now here's my (logical?) counter-argument:

1. The 4080 is better than even a 3090 Ti which was released at an MSRP of 1,999$. People were not as negative towards the 3090 Ti as they are over the 4080.

Thoughts?

People are more criticising nVidia's broken logic of "we made it +70% faster so pay +70% more" which is bad enough for one generation, but when compounded from one generation to the next, ends up completely divorced from reality of the history of computing (improved perf / $ due to progress).

Reality check : If you took an early nVidia GPU say the $125 1999 RivaTNT 2, then applied recent nVidia pricing logic consistently to each generation from its replacement GeForce 256 to now, the fair market value of a modern GPU that's tens of thousands of times faster would be between $1.5m-$3.5m based on either "+70% more $ per +70% fps" or "$125 + 70% annual compound price increases sustained for 20 years". Does that sound sane to you? It's less the GPU itself and more people are bashing nVidia's recent weird attempts to rewrite history of "it's entirely normal for things to get much faster than they get more expensive, it's called 'progress' and applies to every component except for nVidia GPU's". If people were fine with $699 3080's and $1199 4080's, how many generations do you think that can be sustained? $2050 RTX 5080's? $3530 RTX 6080's? $6,050 RTX 7080's? $10,380 RTX 8080's? $17,800 RTX 9080's? $30,500 RTX 10080's? There's obviously going to be a hard crash down to earth somewhere and not a moment too soon...