AMD Lose More Market Share To Nvidia - Now Only Got 18%

[muon]

so you admit it has had issues with accuracy?

Yes once upon a time in 2018 there was a bug.

Not lots of issues. Once there was a bug that was then fixed.

Then in a thread about hardware data in 2021, you call the steam hardware survey numbers which I am quoting, full of crap.

You have googled two other issues (one which isn't an issue and a mistake by the user) and another about game data.

I'm afraid you come out of this looking very bad. What about the JPR numbers which align with the steam hardware survey?

 

[Chuk_Chuk]

As much as the launch stock was disappointing, there was something that I think many people, myself included, missed.

Nvidia were always going to sell more cards than AMD Probably on a 4 to 1 ratio based off the OPs post.

Why does this matter you may ask.
Let's go back to Jan 2020 when we were worrying about WW3 starting with Iran, Australian Bush fires and we all remember what freedom looked like.

You are part of AMDs supply chain team and you need to decide how much production you want for these cards. You are going to use historical data to determine how much capacity you will allocate.

Let's not forget that tooling for each line cost money. Do you spend money building a production line to churn out way more cards per month than you forecast to sell for the once in a blue moon opportunity that your competitor drops the ball on supply or that crypto makes a resurgence? Or do you build a production line to cover what you think you are actually going to sell?

Let's consider the present day from the safety of our socially distanced bedrooms. How long is this insane demand really going to last? How long before Nvidia production line is running at 100%. How much are you going to increase production by if it is even possible?
Didn't AMD get screwed over back in the 290x days. Mining boom crashed and they got left with a bunch of cards they couldn't sell.

I believe they need new masks for each line to produce the different chips so even if they wanted to use some of the console wafers for RDNA2, they would at least need a full set of masks to put on the machines. (are they called masks or templates or something else?)

With this info we can conclude. There will always be more stock of Nvidia Cards this generation. I don't expect the percentage of market share to change significantly this generation (especially since even AMD didn't expect to be this close to Nvidia) but I Expect that now AMD know they are snapping at Nvidia's heals they will probably be more ambitious with their production target for RDNA3.

I heard something about AMD selling more cards now than ever before if true I don't see AMD just doubling their production capacity at best they may give it a 20% bump.

 

[muon]

You are part of AMDs supply chain team and you need to decide how much production you want for these cards. You are going to use historical data to determine how much capacity you will allocate.

Let's not forget that tooling for each line cost money. Do you spend money building a production line to churn out way more cards per month than you forecast to sell for the once in a blue moon opportunity that your competitor drops the ball on supply or that crypto makes a resurgence? Or do you build a production line to cover what you think you are actually going to sell?

I agree that AMD were never going to meet the unexpected demand. I guess the issue is that they are selling very few cards even with this unprecedented demand. It is very reasonable to say it is a TSMC silicon issue, and that is understandable. However, some at AMD did shoot themselves in the foot by laughing at nvidia.

They should really be gaining share and not losing it right now. Ryzen starting clawing back share very quickly.

 

[henson0115]

Yes once upon a time in 2018 there was a bug.

Not lots of issues. Once there was a bug that was then fixed.

Then in a thread about hardware data in 2021, you call the steam hardware survey numbers which I am quoting, full of crap.

You have googled two other issues (one which isn't an issue and a mistake by the user) and another about game data.

I'm afraid you come out of this looking very bad. What about the JPR numbers which align with the steam hardware survey?

no 2016/17/18, via all the links. there's hundreds more again on google. yes i posted 1 mistaken 1 by a user so what? 1 bug is enough to question the voracity of data. who's to know how many more bugs there are.
all you are doing is deflecting, i do call it full of crap, thats my opinion and who are you to tell me i cant say it isnt? get over yourself, again perhaps grow up and maybe take it less personal (its just hardware).

 

[muon]

no 2016/17/18, via all the links. there's hundreds more again on google. yes i posted 1 mistaken 1 by a user so what? 1 bug is enough to question the voracity of data. who's to know how many more bugs there are.
all you are doing is deflecting, i do call it full of crap, thats my opinion and who are you to tell me i cant say it is? get over yourself, again perhaps grow up.

Do you read your own links? Because I did.

The 2016 one is about games.

All the others are about the same issue. Please read them next time.

 

[henson0115]

Do you read your own links? Because I did.

The 2016 one is about games.

All the others are about the same issue. Please read them.

and yet still wrong.. wrong is wrong regardless of the data collected.

 

[muon]

and yet still wrong.. wrong is wrong regardless of the data collected.

Lol. Backtracking again.

I thought there were multiple issues with the hardware survey from 2016, 2017 and 2018?

There has been a reported issue once with the Steam Hardware Survey.

You claimed off the bat there had been multiple issues. Now you can't even show them. That stinks of BS from the very beginning.

edit:

Reading about the 2016 game link it's not actually an issue with published steam data. It's something SteamSpy (a third party) is aggregating catalogue information and then presenting.

 

[henson0115]

Lol. Backtracking again.

I thought there were multiple issues with the hardware survey from 2016, 2017 and 2018?

There has been a reported issue once with the Steam Hardware Survey.

You claimed off the bat there had been multiple issues. Now you can't even show them. That stinks of BS from the very beginning.

edit:

Reading about the 2016 game link it's not actually an issue with published steam data. It's something SteamSpy (a third party) is aggregating catalogue information and then presenting.

even if the 1 issue spread across multiple years, its still 1 issue to many when talking about data f millions of users and calls into question the voracity of any further surveys. who's to say how many more bugs there is or isn't. again, there is no grey with you is there?

 

[muon]

even if the 1 issue spread across multiple years, its still 1 issue to many whwne talking about data form millions of users and calls into question the voracity of any further sureys. whos to say how many more bugs there is or isn't. again, there is no grey with you is there?

It's not spread across multiple years.

They are talking about the same issue. Seriously, read your links.

 

[henson0115]

It's not spread across multiple years.

lol ok then. again deflecting.

 

[Gerard]

AMD need TSMC to make their products competitive against Nvidia, also moving to a different foundry and a completely different process node may well have required a very extensive redesign.

Hes also 100% wrong about the "masterstroke" and trying to spin second choice into some work of genius on nvidias part. They wanted tsmc, they tried to play hardball with tsmc, tsmc didn't budge so nvidia had to go with second choice on the lesser node. But of course he tries to spin that like it was some "great idea" when it was nothing but pure chance and nvidias little powerplay falling flat on its face.

Yeah nvidia has better supply for some skus but at vastly increased prices unless you're lucky enough to beat the literal thousands watching telegram and part alert notifications waiting to pounce on an fe card.

I swear the shills will try and spin anything.

 

[muon]

lol ok then. again deflecting.

The 2019 link talks about August 2017. Mentions Valve fixed it.

Then it talks about SHS in general without mentioning any other specific issue.

Seriously, read your own link.

 

[henson0115]

The 2019 link talks about August 2017. Mentions Valve fixed it.

Then it talks about SHS in general without mentioning any other specific issue.

Seriously, read your own link.

https://www.roadtovr.com/valve-tweaks-steam-survey-accurately-count-vr-headsets-pimax/
"Valve tells Road to VR that it’s revamping the way it collects data about VR headsets used on Steam, which should increase the accuracy of the figures in the Survey. The company says that both wireless and less common headsets will now be represented in the data."
"Previously, the company explained, VR headsets were counted by the Survey only if they were plugged in via USB at the time the data was collected. The new method will allow SteamVR to report to the Survey any headsets that have been used by the system in the last month, regardless of whether they are plugged in at the time of data collection."

 

[muon]

https://www.roadtovr.com/valve-tweaks-steam-survey-accurately-count-vr-headsets-pimax/
"Valve tells Road to VR that it’s revamping the way it collects data about VR headsets used on Steam, which should increase the accuracy of the figures in the Survey. The company says that both wireless and less common headsets will now be represented in the data."
"Previously, the company explained, VR headsets were counted by the Survey only if they were plugged in via USB at the time the data was collected. The new method will allow SteamVR to report to the Survey any headsets that have been used by the system in the last month, regardless of whether they are plugged in at the time of data collection."

You are grasping at straws. This isn't a bug.

They then started collecting data for less commonly used wireless VR headsets a year ago.

This isn't an error.

Using this to support your theory the survey is "full of crap" and throw shade on anyone use their numbers is just hilarious.

You also knew none of this when you made your statement.

 

[henson0115]

You are grasping at straws. This isn't a bug.

They then started collecting data for less commonly used wireless VR headsets a year ago.

This isn't an error.

didnt say it was a bug.

 

[muon]

it is full of crap, it has been manipulated (unintentionally) and steam had to fix the issue on multiple occasions, that doesn't mean that its 100% guaranteed to be accurate now, they have found many bugs and issues with it how naïve are you? maybe you should do some research before slinging mud..

im not referring to rdna 2 cards at all, but also ampere.

acting all insulted because i called your little survey crap. grow up maybe?

You are getting really desperate in trying to discredit the survey. I don't get why.

 

[henson0115]

You are getting really desperate in trying to discredit the survey. I don't get why.

lol ok then. you still cant take away the fact that it has been wrong regardless of what you say. again i question the voracity of all future data based on this. you can make of that what you will i really dont care. enjoy your smug black and white life

 

[muon]

lol ok then. you still cant take away the fact that it has been wrong regardless of what you say. again i question the voracity of all future data based on this. you can make of that what you will i really dont care. enjoy your smug black and white life

You've been wrong multiple times in this thread and keep digging yourself a bigger hole. Shall we apply your logic?

Steam Hardware Survey has been reported as wrong once and corrected itself after identifying and publishing the details of the issue.

I'm the one willing to let a 3 year old single issue slide and so am not holding an absolute position like you.

 

[henson0115]

You've been wrong multiple times in this thread and keep digging yourself a bigger hole. Shall we apply your logic?

Steam Hardware Survey has been reported as wrong once and corrected itself after identifying and publishing the details of the issue.

I'm the one willing to let a 3 year old single issue slide and so am not holding an absolute position like you.

using your logic, they fixed it forever and now its 100% accurate despite not including the 50 million+ miners out there excluded from the charts or the fact its programmed by people who cant make mistakes and miss things.
valve has stated itself that it has worked on the accuracy, itself meaning that it wasn't as perfectly accurate as you believe it to be.

absolute position? ive told you that's its my opinion on multiple occasions and you've persisted in telling me im wrong. yes i made a mistake at the very start.

again this is all good fun, but you seem to care about it all way to much! replying back telling me how wrong i am for the past half day merely proves that.

dont you think this thread is derailed enough?

 

[Nexus18]

Steam survey is as good as you are going to get and tbf, it's a pretty good indication given that every pc gamer will have it installed. The ones who are buying solely for mining are not going to be installing steam.....

Also, it's already been stated that up to 80% of amds hardware is going to consoles:

https://www.hardwaretimes.com/amd-p...nm-capacity-allocated-to-console-socs-report/

Then factor in if you are signed up to part alerts to be notified of when stock comes in, nvidia must outweigh amd stock alerts by about 10-1. When I was signed up, 3080/3070 were coming into stock at least 10 times a day where as for amds 6800(xt), you were lucky if you seen 5 items every 2 days....

So it's no surprise AMD is losing market share in the pc gaming scene.

 

[fs123]

I agree that AMD were never going to meet the unexpected demand. I guess the issue is that they are selling very few cards even with this unprecedented demand. It is very reasonable to say it is a TSMC silicon issue, and that is understandable. However, some at AMD did shoot themselves in the foot by laughing at nvidia.

They should really be gaining share and not losing it right now. Ryzen starting clawing back share very quickly.

AMD has console commitments so I suspect most of their production quota is dedicated to PS5/XSX. The demand for Ryzen chips is also very high which they are focusing on. Their gpu's will have to take a backseat for a while since there is no contract forcing them to produce more.

As for Steam survey, While I have owned mainly AMD gpu's, I have only been asked to take it once in about 15 years so not sure how it can be that reliable. Is there a way to take it without having to be offered it?