• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Nvidia really doesn't seem to care about gaming GPUs anymore - Gaming GPU sales no longer reported

Draugoth

Gold Member
iyBByNqAxcnWCkwy4otKPS-970-80.jpg.webp
  • Nvidia just announced its Q1 fiscal 2027 results
  • This came with a change in the way GPU sales are reported
  • They won't be detailed separately anymore, but buried in another category —Edge Computing — and there are reasons to be nervous
Tom's Hardware noticed that aside from the record revenue in Q1 fiscal 2027 – which hit a staggering $81 billion – Nvidia is making a change to the way the company reports its financials going forward.

From this quarter and in the future, Nvidia won't separately report sales of client graphics cards, meaning consumer (GeForce) and professional (RTX Pro and others) GPUs.

Article:
Nvidia really doesn't seem to care about gaming GPUs anymore — the company won't even bother to break down graphics sales in its big investor reports
 
I wonder how much money is being generated from the AI enthusiast market these days vs PC gamers. There is some overlap, but those DGX Spark devices start around the MSRP of a 5090
 
Last edited:
The current estimate for Blackwell and Rubin is over $1 TRILLION sell through. Gaming GPUs are literally Jensen's hobby at this point.

Also, shit is going to get worse as Rubin is projected to be 2x Blackwell in performance and will be even more memory hungry.


k25zUUmdrWDIbcmp.gif
 
They will come crawling back. Watch!!

They are never coming back. The kind of money Nvidia needs to keep having record breaking profits is such that the consumer market can't provide it. The only way Nvidia and co can make record breaking profits at this point is to keep making record breaking deals with business customers.

Once Chinese GPUs get good enough prices will come down and they will be even less inclined to sell DIY GPUs.
 
I wish Radeon could compete like they did in the ATI days
Resurrect 3DFX... And bring back ATI as a moniker (it should of never been retired). We could do for a revamped modern voodoo line, or a Matrox mystique , power vr kyro ii. (Im probably showing my age)

Back then those companies all had different approaches to rendering . Of course we had multiple graphics apis to choose from and not every game was going through one engine.

Seriously what is stopping a company from competing? Why aren't there other companies making ram too? There is a niche that isn't being filed someone would do well stepping up. I get it things aren't as simple as back then but the supply and demand concept still is.
 
Last edited:
I'm just sad we'll be wasting years and years of graphical and technological advancements, because of this chase for big bucks. Once the VC money dries up, we'll see who will be left standing.
 
PS6 might be the last consumer at home gaming device. Make sure you buy two and put one away in storage for when the first one wears out you have a replacement. Because when it's gone, it's GONE.
For everyone else, pay by the hour for online gaming access.
 
I wonder how much money is being generated from the AI enthusiast market these days vs PC gamers. There is some overlap, but those DGX Spark devices start around the MSRP of a 5090
DGX Spark is widely considered to be a worthless crap.

Purely from GPU / VRAM pespective, it's mostly these things that are widely accepted as go-to choices:

- RTX 3090 (with 2 of them, you can run some strong local models at 48GB VRAM)
- 5090, 4090, 4090 48GB mod (obviously expensive af to get more than 1 card)
- 7900 XTX (using llama.cpp Vulkan instead of CUDA)
- RTX Pro 6000 Blackwell (one 96GB card is enough, super expensive)
- stacking a few RTX 5060 Ti 16GB cards (slower than RTX 3090 by 1.5 to 2 times?)

While all the gamers whine about "AI enthusiasts" stealing their GPUs, those "AI enthusiasts" largely don't give a shit about anything other than a bunch of used RTX 3090 cards (at least when they aren't into image/video generation; for that stuff, 5090 would be quite significantly better).
 
Last edited:
Bro, in the States, they are starting to use imminent domain to seize properties destined for data center use.
It's not a wide spread thing atm but it's a growing concern of mine.
Fucking crazy. Really starting to look like a fork in the road between Star Trek post-scarcity society and SkyNet / Matrix style demise of mankind ... I know where I'm putting my money!
 
DGX Spark is widely considered to be a worthless crap.

Purely from GPU / VRAM pespective, it's mostly these things that are widely accepted as go-to choices:

- RTX 3090 (with 2 of them, you can run some strong local models at 48GB VRAM)
- 5090, 4090, 4090 48GB mod (obviously expensive af to get more than 1 card)
- 7900 XTX (using llama.cpp Vulkan instead of CUDA)
- RTX Pro 6000 Blackwell (one 96GB card is enough, super expensive)
- stacking a few RTX 5060 Ti 16GB cards (slower than RTX 3090 by 1.5 to 2 times?)

While all the gamers whine about "AI enthusiasts" stealing their GPUs, those "AI enthusiasts" largely don't give a shit about anything other than a bunch of used RTX 3090 cards (at least when they aren't into image/video generation; for that stuff, 5090 would be quite significantly better).

Why is it considered crap with all that memory? I get there are alternatives, but very few of them add up to 128GB. Not saying you are wrong. Just trying to understand.

I've dabbled a bit at home running local models, but I'm mostly using services at work. But yeah, the "enthusiasts" are not impacting the GPU market. Definitely agree there.

Every few months: "NVIDIA doesn't care about gaming."
Every couple of years: NVIDIA releases a gaming GPU that blows everything else out of the water.

People often make the mistake of thinking that if a source of revenue in one segment is small compared to another, then the company "doesn't care" about that revenue. That is simply not the case. Looking at the graph that Aenima Aenima provided, the gaming segment hasn't decreased. Nearly $3 billion in revenue isn't something any company "doesn't care" about.
 
Last edited:
I remember Jensen telling the story of when he was trying to get any investor to make nvidia, during the early 90s. And when they had months left of investor money and the company was almost going under.
4 decades later and it's the most valuable company in the world, probably in human history. What an incredible rise.

PS: Just a shame that they have abandoned gamers and the AI boom they created is fucking gamers in the ass.
 
Last edited:
Every few months: "NVIDIA doesn't care about gaming."
Every couple of years: NVIDIA releases a gaming GPU that blows everything else out of the water.
Yep.

Nvidia: Improving drivers, DLSS, frame gen, Path tracing, Switch 2, etc...
Some people: Nvidia doesn't care about gaming anymore...
 
They might not put it front and center anymore, but even that teeny tiny itsy bitsy bit of the available budget that Nvidia puts into its gaming related developments for sure makes whatever Sony and AMD can put behind their efforts look like pocket change.
That's how dramatic the size differences have become.
They'll not drop their gaming tech development, but it will be pretty much just for PR as long as the AI ramp up is still ongoing and the margins will surely keep rising too.
 
Last edited:
I have doubts If even rtx 6000 will happen for consumers. China need to accelerate their gpus or will be dependent of cloud.

AMD will have something but probably only under 300w.
 
DGX Spark is widely considered to be a worthless crap.

Purely from GPU / VRAM pespective, it's mostly these things that are widely accepted as go-to choices:

- RTX 3090 (with 2 of them, you can run some strong local models at 48GB VRAM)
- 5090, 4090, 4090 48GB mod (obviously expensive af to get more than 1 card)
- 7900 XTX (using llama.cpp Vulkan instead of CUDA)
- RTX Pro 6000 Blackwell (one 96GB card is enough, super expensive)
- stacking a few RTX 5060 Ti 16GB cards (slower than RTX 3090 by 1.5 to 2 times?)

While all the gamers whine about "AI enthusiasts" stealing their GPUs, those "AI enthusiasts" largely don't give a shit about anything other than a bunch of used RTX 3090 cards (at least when they aren't into image/video generation; for that stuff, 5090 would be quite significantly better).

number-one-bullshit-khabib-nurmagomedov.gif


The DGX Spark is 140W(double for some though honestly pointless) all in with 128GB of memory.
Every single GPU you listed is no match for that even if you paired it with a 12100.

An undervolted 5090D2 with 64GB of VRAM is still worse overall than a DGXS and ironically you could(dont know about now) get them at pretty much the same price.


If electricity/effeciency is no issue.....which makes me assume money is no issue then a Pro 6000B x (\(\infty \)
But if you are training and or crunching out models, the DGX Spark pretty much trumps any consumer level GPU for models that would need that kinda memory.
The CUDA core count matters less and less as your model gets heavier and heavier, it starts to make sense to get lower power high GB devices.
 
Last edited:
Every few months: "NVIDIA doesn't care about gaming."
Every couple of years: NVIDIA releases a gaming GPU that blows everything else out of the water.
It's because people see it only makes up a small part of revenue that they seem to ignore the fact that the small part is still tens of billions of dollars yearly revenue.

Last year was 18 billion on gaming GPUs alone. Still a fraction of datacenter sales, but still a staggering amount of money.
 
Top Bottom