NVIDIA GeForce RTX 5080 Founders Edition

[Bwaze]

I'm wondering if the MSRP cards will even be stocked this time take gigabyte for example they have one MSRP card and then it jumps to 1200 for example and how long will it take to log onto newegg/Bestbuy/Amazon and easily buy at msrp pre tax ofc

AIB partners will simply make Reference model and other cheaper SKUs in extremely low numbers, and use their parts to make higher end cards that have larger margin of profit. They have already said that Nvidia made MSRP too low, and number of allocated chips also very low.

A perfect combination for scalping season.

 

[rv8000]

Apart from cryptomining times I always saw MSRP models here in Europe - that is MSRP + tax + maker's tax - you can't really expect Asus Reference card for a MSRP, even though they do promise it at first. But various Zotac, Palit, INNO3D, PNY, Gainward could always be found.

But now is the exception, at least in high end, AI acceleration is the new cryptomining, so higher end model cards were overpriced - for instance RTX 4090 were listed for 400 - 1000 EUR over their MSRP. I'm sure that will carry over to RTX 5090, perhaps even to RTX 5080 now that it has 16 GB of RAM.

I don’t check often, but the majority of bottom tier 4080s (windforce, tuf, ventus, pny oc etc…) were and generally have been $1100 while in stock. Even at launch I don’t remember seeing a card in stock below that price. FE being sold by Nvidia and drip fed to Best Buy once in a blue moon is moot. Remember, selling at MSRP is akin to “Charity” now for AIBs. I have an expert, and even at $1149 it was one of the cheaper cards around release.

 

[calhau]

This is so funny.
When the 4080 was announced people lost their minds because Nvidia was trying to sell the heavily cut down 12Gb model for 900$, the outcry was so great that they made it the 4070ti and 800$.
Now they do the same thing but don't release a model in between the 5090 and this, and no one bats an eye.
Nvidia learned from that mistake, consumers didn't, hell Nvidia can even charge 100$ extra over the MSRP of the "4080 12Gb" (200$ over the MSRP of 4070ti) and everyone is happy.

Core wise:

5080 is 49% of a 5090

4070ti is 47% of a 4090

3070 is 55% of a 3090ti

2070 is 50% of a titan rtx

1070 is 54% of a titan

Lesson of the day: treat consumers as if they were brain dead to maximise profit.

 

[LastDudeALive]

AMD has fumbled some significant aspect of every single GPU launch for 15 years. None of their recent moves behind the scenes inspire confidence they know what they are doing with RDNA 4. Expect cards with a mild price/performance advantage in raster vs the Nvidia competition, meaningfully worse RT performance, and a dramatically worse software feature set. They seem super content with lousy second-rate products which fit that description, and I imagine there's some incompetent exec over there with constant shocked Pikachu face that they are permanently stuck at like 10% market share.

Yep. Fumbling ray tracing and upscaling killed their discrete GPU line. I'll even give them a pass for the RX 6000 series, as they were already in the process of developing RDNA 2 when DXR and DX12 Ultimate were released, and they get some credit for putting it in anyway, even if it sucked. They had a big chance with RX 7000 coming down from the crypto and supply chain shortages, everyone was ready for new GPUs. If they'd had competitive ray tracing and AI solution, even if they'd made it RX 7000 exclusive, they could have had a chance. But the RT still sucked. Ray tracing has been in common use since Cars (2006), surely they could find some software developers from Hollywood or something to help them figure out how it works. But instead they're basically an Nvidia Cargo Cult.

 

[medi01]

And he said 5070>4090.

Logically, 5070 > 5080

Poor AMD.

 

[Bwaze]

I don’t check often, but the majority of bottom tier 4080s (windforce, tuf, ventus, pny oc etc…) were and generally have been $1100 while in stock. Even at launch I don’t remember seeing a card in stock below that price. FE being sold by Nvidia and drip fed to Best Buy once in a blue moon is moot. Remember, selling at MSRP is akin to “Charity” now for AIBs. I have an expert, and even at $1149 it was one of the cheaper cards around release.

Interesting, I was following RTX 4080 Super here in EU, and I think most cheap models dropped bellow MSRP after two months - I think MSRP + tax in Germany was about 1110 EUR, and a lot of cards were available quite a bit below that, some even as low as 950 EUR for short periods, but many around 1000 EUR.

 

[bitsandboots]

I really don't think ray tracing and upscaling was what killed AMD's GPUs.
The reputation damage was already done from years of software issues. Nvidia became known as the premium brand and AMD did very little to try to change that reputation, and eventually got tired of trying to undercut nvidia, and now they have been caught unprepared on ever growing popularity of GPGPU tasks with ROCm being a too little too late thing. Though I do hope its NOT too late, and is just the beginning of a much longer sign of competition!

 

[Bwaze]

I really don't think ray tracing and upscaling was what killed AMD's GPUs.
The reputation damage was already done from years of software issues. Nvidia became known as the premium brand and AMD did very little to try to change that reputation, and eventually got tired of trying to undercut nvidia, and now they have been caught unprepared on ever growing popularity of GPGPU tasks with ROCm being a too little too late thing. Though I do hope its NOT too late, and is just the beginning of a much longer sign of competition!

Does the reason even matter? I think AMD simply chose their battle - they are limited on how many chips they can order from TSMC, top end GPU processors are large, and they have much higher profits in server CPUs - of course that reasoning didn't count the surge in demand for AI acceleration GPUs, but it's a bit late now to change the course...

 

[medi01]

Fumbling ray tracing and upscaling killed their discrete GPU line.

Did you check actual benchmarks, cough? On this very site, cough?
I guess not. Let me help you:

 

[bitsandboots]

Does the reason even matter? I think AMD simply chose their battle - they are limited on how many chips they can order from TSMC, top end GPU processors are large, and they have much higher profits in server CPUs - of course that reasoning didn't count the surge in demand for AI acceleration GPUs, but it's a bit late now to change the course...

Well I don't expect the 9070 to be that bad considering nvidia is at a standstill this generation. But, the way AMD is behaving doesn't give me confidence that they're being strategic vs just retreating. High end isn't everything, and they're doing great in integrated, but they've just vanished from laptops otherwise and that's also curious.

 

[mkppo]

Genuinely curious why the 5080 isn't faster in RT compared to 4080? It's actually less of an uplift compared to the already terrible raster uplift. This is actually a bigger talking point because hey, isn't nvidia all about them RT feels?

People claim nvidia's RT superiority like AMD doesn't exist and IIRC the 7900XT was some 12% slower in RT vs 4070Ti but for some reason the performance was so little they should go bankrupt. Then there's the 4060Ti 8GB which is slower in RT than 7700XT. Now here we are with a full 7% gen on gen RT improvement. Wait, let me guess, RT is also like raster now and 'cant really be improved much since it's already so fast'?

I guess with zero RT performance improvements this gen, I present to you:

Geforce X 4080 Super GDDR7 Edition.

And this thread was sure entertaining, double standards and all.

 

[medi01]

On top of it, 6 years into "hardware RT", of the 37 modern games with that gimmick, we have "wopping" 30% in which it clearly improves visuals.

And on top of that, most users keep RT gimmick off, as the most recognizable RT feature is... wait for it... major FPS drop.

I think AMD simply chose their battle - they are limited on how many chips they can order from TSMC, top end GPU processors are large

7900 XTX Ti does not make sense financially.

I haven't seen anyone being excited about 5090, that was buffed the most in the 5000 gen.
It's so much worse below that.

It's a lackluster generation. FSR4 improvements will matter more than anything else in case of AMD.
150+ "AI notebook" design wins are also quite notable.

 

[W1zzard]

@W1zzard I just noticed that there's a hypothetical 5080 for $800 in the performance per dollar charts. Is that an error?

View attachment 382367

it's for what-if

 

[bitsandboots]

And on top of that, most users keep RT gimmick off, as the most recognizable RT feature is... wait for it... major FPS drop.

Love that HW unboxed did actual research on the state of RT. Since in a game like hogwarts legacy, the most obvious thing it did was make it run out of vram, making the game suffer big time.

 

[3valatzy]

Genuinely curious why the 5080 isn't faster in RT compared to 4080? It's actually less of an uplift compared to the already terrible raster uplift. This is actually a bigger talking point because hey, isn't nvidia all about them RT feels?

Because it's the same card in a new shiny retail package.

 

[csendesmark]

Well, for this money I would expect more

 

[Hxx]

This release is more boring than a wet wipe . Thanks nvidia you greedy green goblin!

 

[RayneYoruka]

The AIB models over here range from 1200 to 1800€. Being 1600 the MSI Gaming model I've been looking for. Perhaps I should wait a year or so before I upgrade.

 

[JustBenching]

On top of it, 6 years into "hardware RT", of the 37 modern games with that gimmick, we have "wopping" 30% in which it clearly improves visuals.

How can RT be a gimmick while at the same time it clearly improves visuals, as you just said so yourself?

Now let me make a guess. How many of those games that don't look better with RT ON in that list of yours are sponsored by the company that is clearly behind in RT performance - and so they have a vested interested in keeping RT in games to a minimum so they don't get exposed for their lackluster RT performance?

So yes, you are correct, AMD is holding back gaming due to their own incompetence.

 

[BigMack70]

This is so funny.
When the 4080 was announced people lost their minds because Nvidia was trying to sell the heavily cut down 12Gb model for 900$, the outcry was so great that they made it the 4070ti and 800$.
Now they do the same thing but don't release a model in between the 5090 and this, and no one bats an eye.
Nvidia learned from that mistake, consumers didn't, hell Nvidia can even charge 100$ extra over the MSRP of the "4080 12Gb" (200$ over the MSRP of 4070ti) and everyone is happy.

Core wise:

5080 is 49% of a 5090

4070ti is 47% of a 4090

3070 is 55% of a 3090ti

2070 is 50% of a titan rtx

1070 is 54% of a titan

Lesson of the day: treat consumers as if they were brain dead to maximise profit.

Getting mad at consumers/customers here is a really weird take. Monopolies have nearly complete power over their customer base. The 4080 12GB / 4070 Ti had competition from AMD. The 5080 does not. This isn't a "why are customers so braindead" moment, it's a "monopolies suck and we need AMD or someone else to become competitive" moment.

Nvidia has zero competition on the high end, and the AI bubble means they will sell cards at whatever price Nvidia wants even if gamers completely ignore them. Blaming consumers for that is a little... odd.

 

[medi01]

How can RT be a gimmick while at the same time it clearly improves visuals, as you just said so yourself?

I'm afraid, you got the "6 years into RT, FPS drop remains the most notable feature of the RT gimmick, in 70% of the 'whopping' 37 games, it barely improves anything, sometimes makes things worse" wrong.

 

[JustBenching]

I'm afraid, you got the "6 years into RT, FPS drop remains the most notable feature of the RT gimmick, in 70% of the 'whopping' 37 games, it barely improves anything, sometimes makes things worse" wrong.

In the link you just posted it says "transforms visuals significantly".

 

[LastDudeALive]

Genuinely curious why the 5080 isn't faster in RT compared to 4080? It's actually less of an uplift compared to the already terrible raster uplift. This is actually a bigger talking point because hey, isn't nvidia all about them RT feels?

People claim nvidia's RT superiority like AMD doesn't exist and IIRC the 7900XT was some 12% slower in RT vs 4070Ti but for some reason the performance was so little they should go bankrupt. Then there's the 4060Ti 8GB which is slower in RT than 7700XT. Now here we are with a full 7% gen on gen RT improvement. Wait, let me guess, RT is also like raster now and 'cant really be improved much since it's already so fast'?

I guess with zero RT performance improvements this gen, I present to you:

Geforce X 4080 Super GDDR7 Edition.

And this thread was sure entertaining, double standards and all.

The RT average is dragged down by Doom Eternal (identical performance, probably too lightweight to have any benefit from better cores and more VRAM bandwidth), and Elden Ring, which has a massive regression from the 5090 and 5080, obviously a driver bug with the game or uncapped fps mod.

 

[N3M3515]

Very interesting the 4080 Super has 88% of the 5080's performance at 4k, but 90% at 4k ray tracing. I would have expected the ray tracing performance to have a bigger gap than the regular performance, due to improved RT cores and more memory bandwidth from GDDR7.

That's because nvidia never actually improves RT cores, they just "add moar"
It's been like that since the first gen of RTX gpus.

 

[mkppo]

On top of it, 6 years into "hardware RT", of the 37 modern games with that gimmick, we have "whpping" 30% in which it clearly improves visuals.

In that list, in the generally better category there's CP2077 with RT Ultra but I prefer mine with RT off. Less bloom, sharper image. Also much higher performance. If that's what I feel about the supposedly 'great for RT' title, I wonder how i'll feel firing up Dead Space 2. It'll be like having a setting that says 'enable this for 40% lower FPS and possibly artifacts'.

I honestly don't get it. Maybe i'm old. But just when I thought it's probably just me, HUB's review comes along and mostly mirrors my experience.

RT has good promise, we're just not there yet and these forced implementations usually just suck.

The RT average is dragged down by Doom Eternal (identical performance, probably too lightweight to have any benefit from better cores and more VRAM bandwidth), and Elden Ring, which has a massive regression from the 5090 and 5080, obviously a driver bug with the game or uncapped fps mod.

I saw Elden Ring and even there, the performance difference between 5080 and 4080 isn't much. Let's say we give it an extra 15%. It'll still barely change the overall 4080>5080 RT performance uplift which seems negligible.