GPU prices — is the worst behind us ?

“The retail price of the anniversary version of the GeForce RTX 5090 has reached $12,000.”​

View attachment 831151
https://www.reddit.com/r/pcmasterrace/comments/1x56bpf/12k_now_at_microcenter/
Eh, just for comparison.
Screenshot_20261015_152619_Brave.jpg
 
Well, I picked up a 5070Ti a few weeks ago. I would have waited for the next gen of cards but at this point I'm not convinced that there will ever be a next gen of consumer gaming GPUs (AI/rampocalypse, latest 5090 news, what's even the business case at this point?).

Good news for everyone else though: knowing my luck prices will now start to come down since I just bought something. You're all welcome! If only.
 
NIVIDA has a even bigger total marketshare on the AI market:
https://carboncredits.com/nvidia-co...n-2025-and-reveals-next-gen-ai-supercomputer/

AMD is a 4%:
View attachment 830434

Wishful thinking doesn't change reality.

We have only NVDIA GPU's in our datacenters and 0 AMD GPU's and that is not chagning as long I am have the power to decide.
Partly because the software is supporting CUDA, not AMD and partly because AMD is a bad performance/$ and partly because AMD lack features in the same way they do on the desktop.

Again the data doesn't agree with the far to vocal minority on forums.

Buddy, this report is from 2024, not exactly reflecting current numbers as more have opted for AMD Helios racks versus 2024 when they didn't have the best options available yet. Get some newer data and not from a company trying to sell data analytic services...
 
If you count APU and IGP, I am talking GPU's, stop being stupid, because in the next moment you talk GPU's and not shitty integrated graphics.
GPU's are like this:
View attachment 831324

Nobody cares for Intel Xe3 (4-core Xe³ graphics package with 512 shaders) or AMD Radeon 860M (8 CU's)
But nice of you once again to display your dishonesty.

And if you ignore FPS and image quality you might have had a point, but in reality, no:

View: https://www.youtube.com/watch?v=SlRpJ553RzE

There is a reason their GPU sales have plummet to 8% flat for many quarters now:
NVIDIA performance > AMD performance
DLSS > FSR
Ray Reconstruction > Ray Regeneration
NVIDIA FG > AMD FG
DLSS 5 > [AMD has Nothing]
RTX Remix > [AMD has Nothing]
CUDA > ROCm
NVIDIA Broadcast AI> [AMD has Nothing]
OptiX > OpenCL/HIP
NVENC > AMD Encoder
NVIDIA Reflex > AMD anti-lag
NVIDIA Studio Ecosystem > [AMD has Nothing]

You keep telling yourself that reality is not like it is, it only makes you part of a cult.

For no one caring, Intel GPU wise has dominance, just not what you care about. The world obviously differs. Now no one cares about your 4090 if they are shopping for a card less than $800. Even Jensen says a 5070 is faster than a 4090. So why compare a 4090 against AMD current generation? Anything game wise graphics card, SUB $850, AMD has clear leadership and for most the better gaming card to get, in performance, in price, availability and I say in general, a more decent community is AMD.
 
For no one caring, Intel GPU wise has dominance, just not what you care about. The world obviously differs. Now no one cares about your 4090 if they are shopping for a card less than $800. Even Jensen says a 5070 is faster than a 4090. So why compare a 4090 against AMD current generation? Anything game wise graphics card, SUB $850, AMD has clear leadership and for most the better gaming card to get, in performance, in price, availability and I say in general, a more decent community is AMD.
Again, yo can mix and match IGP/APU's and GPU's in your "argumentation" like you do.
Facts:
- AMD has dropped to 8% of dGPU slaes for close to 2 years now. Consumers have rejected their dGPU's, despite your crying they do NOT sell as 90% buy NVIDIA, no matter the price point.
- My Intel arc 140T is part of your numbers, but unsued, as i disabled it and use the RTX Pro 1000 exclusively. You are grying to count all the CPU with IGP/APU's not being used for gaming since people adda dGPU tou their system.

So keep crying mr. 8%, reality does not care for your hurt feelings or false claims 8% = 8%.
 
AMD was a little late to the game compared to Nvidia, but I also looked up the second pie chart that you posted and they claimed 5-7% for AMD. A 6GW commitment from Meta is a sizable amount of AI hardware that will change that percentage. I know, you will poo poo AMD any chance that you get, I got your one trick pony show figured out already...

1792154028360.png


https://axis-intelligence.com/amd-statistics/
 
Last edited:
AMD was a little late to the game compared to Nvidia, but I also looked up the second pie chart that you posted and they claimed 5-7% for AMD. A 6GW commitment from Meta is a sizable amount of AI hardware that will change that percentage. I know, you will poo poo AMD any chance that you get, I got your one trick pony show figured out already...
This is a bit of a strange conversation, Meta has something like 12 GW of nvidia installed or commited compute has well.

Meta deal is a superbe one, but it represent 2-3% of the 200-250 GW of committement around the world in that space and its split over 5 years ramp up, it does not mean that it will change that 5-7% figure in a significant way because Meta has been all along AMD biggest customer pretty much and the reason they had a nice 5-7% market share, that just a normal extension of their commitement.

Meta has been of the biggest AMD customer/future owner (they could become an almost 10% owner, depending on how much dilution happen and own much of AMD become OpenAI).

Amd is ramping extremelly fast but so does Nvidia, last 2 years AMD had spectacular growth, but to gain market share in a significant way in that space, you need to be growing faster than Nvidia which had a +116% pace YoY in their last quarter in the datacenter division, you could be growing at never seen before for that scale 70-90% like AMD (or even record breaking 107%...) is and still loose ground versus them....
 
Last edited:
This is a bit of a strange conversation, Meta has something like 12 GW of nvidia installed or commited compute has well.

Meta deal is a superbe one, but it represent 2-3% of the 200-250 GW of committement around the world in that space and its split over 5 years ramp up, it does not mean that it will change that 5-7% figure in a significant way because Meta has been all along AMD biggest customer pretty much and the reasont they had a nice 5-7% market share, that just a normal extension of their commitement.

Meta has been of the biggest AMD customer/future owner (they could become an almost 10% owner, depending on how much dilution happen and own much of AMD become OpenAI).

Amd is ramping extremelly fast but so does Nvidia, last 2 years AMD had spectacular growth, but to gain market share in a significant way in that space, you need to be growing faster than Nvidia, their latest quarter they had a +116% pace over the one of the year before in their datacenter division, you could be growing at never seen before for that scale 70-90% like AMD (or even record breaking 107%...) is and still loose ground versus them....
AMD didn't have rack system comparable to Nvidia previously, but they have their Helios system after having acquired a few companies and some partnerships to put that packaged system together. They are in a much better place to compete now than ever before, but they have a lot of work ahead of them to keep clawing marketshare. Thanks for the engaging post, much preferred over a certain other member that could take some pointers for future engagement.
 
AMD didn't have rack system comparable to Nvidia previously, but they have their Helios system after having acquired a few companies and some partnerships to put that packaged system together. They are in a much better place to compete now than ever before, but they have a lot of work ahead of them to keep clawing marketshare. Thanks for the engaging post, much preferred over a certain other member that could take some pointers for future engagement.
Their systems are still lacking behind, when you attended vendor briefings, it is crystal clears than AMD are generations behind and the gap keep expanding.
 
On that note, the least expensive 5080s on Newegg jumped from $1,6xx to $1,800 this week. I expect they will be at $2,000 by the end of this month.
Newegg, B&H, and Amazon have very few 5080 models in stock. There's a $1,900 PNY at B&H, but everything I'm seeing at Newegg and Amazon is $2,000+. Didn't even take until the end of the month.

Also noting the cheapest 5070 Ti models that were $1,100 are now $1,200+.

Wouldn't be surprised if the 5080 24GB models that are supposed to launch in Q1 2027 start at $2,400.
 
Even 5060 Ti's are stupid priced... I guess if we are answering the question "gpu prices - is the worst behind us?"... that would be a hell no.. not even close..
Yah, and this thread was started over 5 years ago. Can you believe that. I think that these insane GPU prices are the new normal. :bigtears: Pretty soon, if you say you have just built your own PC from parts, that's an oblique way of saying, "Hey I'm loaded!" :eek:
 
Even AMD graphics cards are up (which was a no-brainer that it was going to eventually happen). $829.99 for the cheapest RX 9070 XT 16GB on Amazon ($806.70 cheapest used on eBay). $699.99 for a RX 9070 16GB ($679.99 cheapest used on eBay).
 

“NVIDIA GeForce RTX 5080 & RTX 5070 Ti 16 GB Graphics Cards Witness A Huge Price Increase of Up to $300, As Market Prepares For An Even More Expensive 24 GB 5080​


NVIDIA-GeForce-RTX-50-GPUs-1920x1080-1-728x410.jpg


NVIDIA's partners and retailers in China are raising the prices of GeForce RTX 5080 and RTX 5070 Ti 16 GB graphics cards.

NVIDIA GeForce RTX 5080 Gets A $300 Price Bump While RTX 5070 Ti Gets A $100 Bump in China​

More price increases are on their way for GeForce RTX 5080 and RTX 5070 Ti graphics cards. In China, the latest price bumps are already in effect, with the GeForce RTX 5080 seeing a massive bump in spot prices.”

Source; https://wccftech.com/nvidia-geforce...16-gb-huge-price-increase-24-gb-5080-inbound/
 

“NVIDIA Eyes Software Revenue With a $2,000 per GPU Yearly Figure From Huang​

by DrUsmanSaleem Today, 11:21 Discuss (10 Comments)
NVIDIA CEO Jensen Huang has put a rough price on the software side of the company's AI business. During a Q&A session at GTC Berlin last week, Huang said software licensing will become a growing part of NVIDIA's business, and that it costs about $2,000 per GPU per year to run all the NVIDIA software that powers agents in AI factories. The Register was first to report it, and Huang seems to have offered it as an example rather than a formal price.

The number is a bit confusing, since NVIDIA already sells its AI Enterprise suite on a per-GPU basis. The company's own licensing guide lists it at $4,500 per GPU for a one-year subscription with support, or $18,000 per GPU for five years. NVIDIA hasn't said whether Huang's $2,000 figure covers the same software, a narrower agent-focused bundle, or a future licensing model, so it probably shouldn't be read as a price cut. There's also no word on timing.“
 

“NVIDIA GeForce RTX 5080 & RTX 5070 Ti 16 GB Graphics Cards Witness A Huge Price Increase of Up to $300, As Market Prepares For An Even More Expensive 24 GB 5080​

View attachment 832663

NVIDIA's partners and retailers in China are raising the prices of GeForce RTX 5080 and RTX 5070 Ti 16 GB graphics cards.

NVIDIA GeForce RTX 5080 Gets A $300 Price Bump While RTX 5070 Ti Gets A $100 Bump in China​

More price increases are on their way for GeForce RTX 5080 and RTX 5070 Ti graphics cards. In China, the latest price bumps are already in effect, with the GeForce RTX 5080 seeing a massive bump in spot prices.”

Source; https://wccftech.com/nvidia-geforce...16-gb-huge-price-increase-24-gb-5080-inbound/
I’m glad I’m set with my fairly recent PC builds that should last me until I’ve basically lost interest in this once great hobby that has been ruined by corporate greed and AI.
 
Our only hope is AI to fail, and datacenters selling RTX 5090's for $100 a piece...
That's the big problem back in the day Data centers used stuff that we could reuse fairly well and put windows on it and slap a gpu......now the GPUs and everything else is really only good for running AI stuff form what I gathered.
 
That's the big problem back in the day Data centers used stuff that we could reuse fairly well and put windows on it and slap a gpu......now the GPUs and everything else is really only good for running AI stuff form what I gathered.
I do not feel sorry about building 8 x GPU servers.
It is part of my job.
It pays well.

The doom and gloom is funny considering less than 1% of gamers buy 5090's.
 
Last edited:
Doom and gloom is because the 5080 was $1000 and now it's $2500. The 5080 should not have anything to do with AI but here we are.
I've seen prebuilts with a 5080 for $2500-2700 total still as of the last few days... They may rise to that on their own with time but they're not there just yet. Heck, there was a staples deal for $1450 a week or two ago.
 
I've seen prebuilts with a 5080 for $2500-2700 total still as of the last few days... They may rise to that on their own with time but they're not there just yet. Heck, there was a staples deal for $1450 a week or two ago.
It's here today - head over to Micro Center's website. If you're in the market - buy those prebuilts.
 
It's here today - head over to Micro Center's website. If you're in the market - buy those prebuilts.
I'm not in the market, just saying, deal hunt. No reason to pay that kind of price as of now. Personally, my 4080 mobile laptops and a switch 2 with a bunch of exclusive games are keeping me tided over nicely. I only really game on my pc for bf6 and wow forever right now, and that's been the case for awhile :eek:.
 
I'm not in the market, just saying, deal hunt. No reason to pay that kind of price as of now. Personally, my 4080 mobile laptops and a switch 2 with a bunch of exclusive games are keeping me tided over nicely. I only really game on my pc for bf6 and wow forever right now, and that's been the case for awhile :eek:.

The point is - historically Micro Center represents where the market and pricing is going. Just because there’s a handful of retailers who haven’t caught up is luck not reality.

EDIT: see these "deal ends in" for EVERY RTX 5080 / 5070 Ti on Best Buy? That means the price goes up at Midnight CST (if they haven't yet).

https://www.bestbuy.com/site/searchpage.jsp?id=pcat17071&st=RTX 5080
 
Last edited:
The point is - historically Micro Center represents where the market and pricing is going. Just because there’s a handful of retailers who haven’t caught up is luck not reality.

EDIT: see these "deal ends in" for EVERY RTX 5080 / 5070 Ti on Best Buy? That means the price goes up at Midnight CST (if they haven't yet).

https://www.bestbuy.com/site/searchpage.jsp?id=pcat17071&st=RTX 5080
$2,300 for a 5080 Astral is crazy. The Prime 5080 for $1,600 is decent but sold out. What a time to be alive 😂
 
Back
Top