AMD Says It Will No Longer Go after the Flagship Gaming GPU Market as NVIDIA Continues to Dominate with 88% Market Share

[theblackangus]

Nah, I mean AMD. I find it difficult to understand: sell your tech in consoles for.... Some dollars per console? Granted yes its millions of consoles I get that but this is
vs risk a little by cutting margins in pc and making way more competituve products, at like still more than a few bucks, and remain relevant.

They dont build anything, not the chips, not the cards, nothing, just like with consoles.
Strategy would be aiming for production runs relatively short, in slightly older process, see how it goes every 6 months, renhash, produce again so.on, short contracts/ production runs at x number of units so.on.
It doest have to ever be an ' now im broke' risk for them, for going lower margin.
Market share matters, a lot. They shoukd be fighting way harder than what they are.

I think history shows that AMD providing better performance/$ doesn't gain them market share.
The 3050, while costing more was outselling AMD's better performing competition.
I would be nice if AMD could gain market share but them doing that by pricing has failed quite often.

 

[theblackangus]

The difference is if AMD competes with Nvidia their stock dumps billions in value. If AMD looses just barely to Nvidia AMD stock increases in value...

Because Nvidia has grown so large all AMD has to do is jostle the cart and Nvidia **** near dumps value equal to AMD"s entire market cap. Enough of that and Nvidia will have to get serious about pricing to prevent that.

IT's not about winning... it's about perception.

I don't disagree with the scenario you proposed (in spirit), but historically AMD lowering prices doesn't take much of NVidia market share.
Maybe a decade ago was the last time that worked and coincidence they had the best performing card at that time period. I dont think it has happened since with any significant quantity.

Today with NVidias stock price being tied strongly to the AI trend (bubble?) I doubt the gaming revenue would move the needle much at all.
AMD need to hit NVidia with amazing MI3x0 and MI4x0 products and software ecosystem to do that IMHO.

NVidia KNOWs thier price will be volatile but as long as the AI trend continues I really doubt they care about the gaming revenue, unless they lost a significant chunk.

 

[LazyGamer]

I would be nice if AMD could gain market share but them doing that by pricing has failed quite often.

Generally, AMD has to price lower when for the *average gamer Nvidia provides a better product at a given level of raster performance.

AMD does catch up, and for many gamers the difference would be moot; but that's skipping the latest and greatest. Once you get into comparisons of RT, FSR vs. DLSS, any driver shenanigans, and stuff like Nvidia Broadcast that pretty much just works, someone looking into current generation GPUs is only really going to look at AMD for price.

Which means that, as you said above, price is already a deciding factor, thus cutting their prices even more isn't going to move the needle too much.

What AMD needs to do to gain market share is to provide a better product. Unfortunately in the GPU space, unlike in the CPU space (where they have had the advantage of relying on TSMC while Intel's foundry business has foundered a bit over the years), AMD and Nvidia compete for TSMC volume.

 

[theblackangus]

Generally, AMD has to price lower when for the *average gamer Nvidia provides a better product at a given level of raster performance.

This statement has been proven wrong for about the last decade, as it never gets them to a better market share. They just leave money on the table from the share they have.

AMD does need to catch up on marketing themselves and features, hopefully the new guy in charge of unified architecture and the companies recent announcement on the software focus will give their marketing something better to work with.

I would say they DONT need to gain market share in gaming to provide a better product. They need better marketing, better dev relations (which they are currently working), and most importantly a healthy company which they are doing via server CPUs and AI GPUs like the MI series.
Currently having a great MI series will mean more money for better gaming GPU's and features in the long run.

All IMHO of course.

 

[Stoly]

For those of you that witnessed the Cenozoic era, you'll remember there was a time that AMD (ATI) actually had the majority of the market share.

But it's been decades since nvidia has been the dominant player in gaming, I can't remember the last time nvidia had less than 70% market share.

Everyone keeps saying AMD needs to increase its market share, they have been traying to do that for years. They also tried focusing on the mainstream market more than once and it didn't pay out. Remember when AMD had the performance/feature advantage like the 9800 serires vs the 5900, nvidia still outsold AMD.

Even on the CPU side, in spite of beating intel in performance and price for quite a while, it still has significantly less market share than intel which is in its worst position ever, what's even worse, qualcomm might have a better chance to take a huge bite from intel market than AMD.

 

[theblackangus]

For those of you that witnessed the Cenozoic era, you'll remember there was a time that AMD (ATI) actually had the majority of the market share.

But it's been decades since nvidia has been the dominant player in gaming, I can't remember the last time nvidia had less than 70% market share.

Everyone keeps saying AMD needs to increase its market share, they have been traying to do that for years. They also tried focusing on the mainstream market more than once and it didn't pay out. Remember when AMD had the performance/feature advantage like the 9800 serires vs the 5900, nvidia still outsold AMD.

Even on the CPU side, in spite of beating intel in performance and price for quite a while, it still has significantly less market share than intel which is in its worst position ever, what's even worse, qualcomm might have a better chance to take a huge bite from intel market than AMD.

Voodoo forever! (And Billy!)

While they are less than Intel they are doing increasingly better nearly every year for the last few, which is a great trend for them.
Lots of Cloud contracts etc. But they are still in the growing phase here, so I think this trend will continue for a while at least.
Be interesting to see where CPU lies in another 5 years.

 

[Stoly]

Voodoo forever! (And Billy!)

While they are less than Intel they are doing increasingly better nearly every year for the last few, which is a great trend for them.
Lots of Cloud contracts etc. But they are still in the growing phase here, so I think this trend will continue for a while at least.
Be interesting to see where CPU lies in another 5 years.

It may be the alzheimer talking but I don't recall 3dfx being the dominant market share leader.

 

[theblackangus]

It may be the alzheimer talking but I don't recall 3dfx being the dominant market share leader.

Sorry, wasn't implying that. I was cheering on your Cenozoic era comment with something and someone else from that period.

 

[Stoly]

Sorry, wasn't implying that. I was cheering on your Cenozoic era comment with something and someone else from that period.

it's not that, I'm not sure 3dfx was ever the market leader or if it was, it was short lived, even though no one can deny it was the absolute best tech back in the day.

I didn't have a voodoo card back then; it was during the 3dfx fall that both a voodoo2 and a voodoo 3 fell into my hands. It was still awesome.

 

[Uvilla]

Well I mean they have no where to go in GPU or APUs. They cant do top tier, (they could do top tier APUs right now, but wont do that either, they should) and this mediocre "discount' strategy is taking them nowhere.
I dont know how long they did the value strategy mentioned above, what I am saying is hammer and tongs years long strategy of providing better value aggresively in both APU and dGPUs.
Doing so can leave them with significant advancements, doing agressive 6mo cycles would help too, it will give them all kinds of applicable tech, even cheaper AI cards, creative use of older chip tech etc.
Also they should just include all manner of media and 3d acceleration in all cards /apus and call it a day, Only charge for either professional drivers (only difference would be most stable release and support included) or charge for direct customer service for driver support if either makes them money, if not, just give up the professional acceleration distinction and include it in all cards all those tools and call it a day.

Ryzen won and kept winning cause its been hammer and tongs in price/performence, and even performance crown but its taken years and years, and its not like Intel is dead or the minority, but AMD is making money with this sustained attack.

 

[LazyGamer]

They cant do top tier, (they could do top tier APUs right now, but wont do that either, they should)

They can't do much more than they already are; APUs are monolithic, so they likely cost AMD more to build per die than their individual CCDs and IODs used in 'desktop' and server parts.

Further, the real constraint for APU performance is memory bandwidth, and increasing memory bandwidth also has significant die and platform costs. More contacts on the die, more pins for the socket, more traces on the board, more individual memory dies / packages / modules / whatever. If AMD were to dispense with modularity, what we'd get would look a lot like a GPU -- or a console. But the individual cost would be very high and almost certainly wouldn't be available directly to end-users.

From AMDs perspective it just isn't likely to be a money-making venture.

 

[Niner51]

This makes me think back to ATI and how they were competitive with Nvidia back in the day.

 

[Marees]

By a coincidence AMD acquired ATI the same year that Nvidia released cUda
They will release Udna in 2 years

Just 2 decades worth of catching up to do ...

 

[Snowy]

Part of AMD's problem in market share IMO is they have never made the connection between their console GPUs and desktop GPUs. There are tens of millions of people that play games on AMD GPUs every minute of the day and are have a great experience. The only difference is they're playing on an Xbox or PlayStation. AMD has never leveraged this advantage into developer resources or marketing material for desktop GPUs. You would think that when you're the primary hardware maker for the entire console market you could leverage that to your advantage, yet they haven't for whatever reason.

 

[Grimlakin]

Part of AMD's problem in market share IMO is they have never made the connection between their console GPUs and desktop GPUs. There are tens of millions of people that play games on AMD GPUs every minute of the day and are have a great experience. The only difference is they're playing on an Xbox or PlayStation. AMD has never leveraged this advantage into developer resources or marketing material for desktop GPUs. You would think that when you're the primary hardware maker for the entire console market you could leverage that to your advantage, yet they haven't for whatever reason.

Probably a contractual agreement so they don't dilute the console market. It would be small.. but even that is enough of a risk to console makers.