A laptop GPU now leads the Valve Steam survey, AMD CPUs reach all-time high

[Squid Surprise]

Like said, multiple possible reasons. Perhaps most computers that have multiple users have more Nvidia GPU. Or software bug on Steam software. Or...

If (or should I say because) Steam survey does not randomly pick participants, then whole statistic dissussion is totally irrelevant.

In any case, claiming Steam survey = Sales is totally wrong.

So nothing… you’re just grasping at straws…

 

[HardReset]

So nothing… you’re just grasping at straws…

That was more than enough to prove Steam survey is not same as sales.

 

[Squid Surprise]

That was more than enough to prove Steam survey is not same as sales.

But you’re implying that it somehow is biased in Nvidia’s favour!
Not being the same is NOT the same as being pro-Nvidia!

 

[HardReset]

But you’re implying that it somehow is biased in Nvidia’s favour!
Not being the same is NOT the same as being pro-Nvidia!

What Steam survey tells about GPUs, it's very clear that is way different than actual sale figures. Because of that, it's very easy to claim Steam survey is in some way biased towards Nvidia.

I personally have multiple virtual machines and each have different Steam account. I sometimes launch Steam survey on those virtual machines to wreck statistics. Perhaps I'm not only one

 

[Squid Surprise]

What Steam survey tells about GPUs, it's very clear that is way different than actual sale figures. Because of that, it's very easy to claim Steam survey is in some way biased towards Nvidia.

I personally have multiple virtual machines and each have different Steam account. I sometimes launch Steam survey on those virtual machines to wreck statistics. Perhaps I'm not only one

Again, why would this benefit nvidia and not AMD?
Any evidence that Steam somehow counts one nvidia PC as many, yet counts AMD PCs as single entities?

 

[manowar51]

This is why Techspot should stop posting news about this stupid survey. See, Steam survey has Never been about sales. For last 15 years Intel has had more GPU sales than Nvidia and AMD combined. Steam survey says otherwise. Keep dreaming on Steam survey telling something about Sales ""

I disagree actually, I think Techspot should keep publishing articles on this. I find the Steam survey seems to reflect the state of the western gaming market quite well. Yes, it's not direct sales but when a popular product launches, they tend to go to the top of the charts in the months that follows. I actually think it's a very good resource to you use to determine how well dedicated GPUs are selling to gamers. I should also point out that Intel UHD graphics is high on the list. I dont think it's stupid to assume the 5000 series has sold quite well, it sold out everywhere and marked above MSRP where its not. Poorly selling products don't get this, they get price cuts. Much like the Radeon 7000 series did and we know they were particularly un-popular with gamers.

Im curious, do you have a better source? Something that can disprove the popularity of the cards at the top of the steam survey?

 

[HardReset]

Again, why would this benefit nvidia and not AMD?
Any evidence that Steam somehow counts one nvidia PC as many, yet counts AMD PCs as single entities?

You already know that is pretty much impossible to prove. However GPU sales and Steam statistics have significant difference. And that difference has to come from somewhere.

I disagree actually, I think Techspot should keep publishing articles on this. I find the Steam survey seems to reflect the state of the western gaming market quite well. Yes, it's not direct sales but when a popular product launches, they tend to go to the top of the charts in the months that follows. I actually think it's a very good resource to you use to determine how well dedicated GPUs are selling to gamers. I should also point out that Intel UHD graphics is high on the list. I dont think it's stupid to assume the 5000 series has sold quite well, it sold out everywhere and marked above MSRP where its not. Poorly selling products don't get this, they get price cuts. Much like the Radeon 7000 series did and we know they were particularly un-popular with gamers.

Im curious, do you have a better source? Something that can disprove the popularity of the cards at the top of the steam survey?

Western gaming market? From survey, Simplified Chinese and Russian alone (from top 3 only) have combined share of 36.73%. Not THAT much about western market tbh.

Intel has sold more GPUs than AMD and Nvidia combined for over a decade now. So that's not surprising at all. I'd like to see actual statistics about 5000 series sales, what actual statistics we have says that AMD sells much better. Again, if something is sold out. It may be just that amount sold was very low.

For example, Q4/2024 AMD discrete market share (AMD has pretty much displaced crap cards with APUs so AMD discrete cards are for gaming) was up around 7% and Nvidia share was down around 7%. And that was seen on Steam survey, well, it didn't. AMD just got more market share on integrated ones, in other words, AMD shipping more discrete cards means AMD got even less discrete share on Steam survey. That doesn't add up.

Here are actual sale figures: https://wccftech.com/mindfactory-week-18-2025-gpu-sales-stats-rx-9070-xt-is-top-seller/

You can say that is only one seller but those are actual units sold. Steam survey have several forged results, I know because I have made several.

 

[Squid Surprise]

You can say that is only one seller but those are actual units sold. Steam survey have several forged results, I know because I have made several.

I think you are suffering from a fundamental misunderstanding of statistics… whether someone “forges” results on a survey containing millions and millions of people is largely irrelevant. They tend to even themselves out.
You state that there is a discrepancy between sales and the results yet don’t actually have the numbers for the TOTAL sales… so… is there a discrepancy? Or are you just deciding that AMD must be better cause you believe it to be so?

 

[HardReset]

I think you are suffering from a fundamental misunderstanding of statistics… whether someone “forges” results on a survey containing millions and millions of people is largely irrelevant. They tend to even themselves out.

Problem with those "statistics" is that they are flawed from beginning. We already know that Steam survey "randomness" (=who gets survey and who don't) comes from client, not Steam servers. Because you can manually start survey, randomness is flawed.

"Someone" with bot farm can easily forge few million statistics if needed.

So:

1. Steam survey randomness is broken
2. Forging results is pretty easy

And yet you tell how statistics work.

You state that there is a discrepancy between sales and the results yet don’t actually have the numbers for the TOTAL sales… so… is there a discrepancy? Or are you just deciding that AMD must be better cause you believe it to be so?

Of course I don't have total sales. However Available sale figures are different from Steam survey data. Like already said regarding Steam survey, there is no better data available so get over it.

 

[Squid Surprise]

Of course I don't have total sales. However Available sale figures are different from Steam survey data. Like already said regarding Steam survey, there is no better data available so get over it.

So… maybe YOU should get over it? Cause the Steam survey is as accurate as you’re going to get - until you can procure better numbers…

 

[HardReset]

So… maybe YOU should get over it? Cause the Steam survey is as accurate as you’re going to get - until you can procure better numbers…

Well, you said Steam survey sample size is millions. Source? Steam survey shows percentages so they must have counted how many surveys contained GPU x vs total amount of surveys. And latter number is?

 

[Squid Surprise]

Well, you said Steam survey sample size is millions. Source? Steam survey shows percentages so they must have counted how many surveys contained GPU x vs total amount of surveys. And latter number is?

While they don’t release the exact numbers, back in 2021 they had about 132 million users… even assuming that hasn’t increased (it has), that would be more than enough for a statistically significant result - a basic understanding of statistics would tell you this…

 

[HardReset]

While they don’t release the exact numbers, back in 2021 they had about 132 million users… even assuming that hasn’t increased (it has), that would be more than enough for a statistically significant result - a basic understanding of statistics would tell you this…

I'm not interested about users, I'm interested about how many of those "users" (which many are inactive anyway) participate on Steam survey. It's known fact that 1. participation is voluntary and 2. not all users are even asked to participate. You are still far from "statistically significant" amount.

Reason they refuse to tell amount of Steam survey participants is very simple. If we know 1. amount of Steam survey participants 2. how many percent have certain video card, we could easily calculate amount of those video cards participating Steam survey. And if that amount exceeds amount of manufactured video cards using that GPU, then we could immediately conclude Steam survey flawed. No wonder they don't want to take risk and refuse to tell exact numbers.

 

[HardReset]

So… maybe YOU should get over it? Cause the Steam survey is as accurate as you’re going to get - until you can procure better numbers…

OK, here's some hard data for you:

AMD total share is: 17.51%
(Intel has: 7,78%)

Now we exclude AMD integrated solutions:

AMD Radeon Graphics 2.10%
AMD Radeon(TM) Graphics 2.01%
AMD Radeon Vega 8 Graphics 0.39%
AMD Radeon Vega 3 Graphics 0.25%
AMD Radeon (TM) Graphics 0.29%
AMD Radeon 780M Graphics 0.21%

Integrated total: 5.25%

That leaves AMD Discrete share on Steam survey to be around 12%.

But looking at AMD discrete GPU share based on shipments, that number is waaaay too low:

NVIDIA vs. AMD Discrete GPU Market Share (2010 to 2025)

Here is the NVIDIA vs. AMD discrete market share in percentage from 2010 to 2024. NVIDIA has performed extremely well in the last decade.

pcviewed.com

Also taking into account that Nvidia low end trash cards like GT710 are discrete ones (GT710 does even make it to the list anywhere) whereas AMD low end trash are integrated, we can safely conclude Steam survey does not reflect on real GPU sales.

 

[Squid Surprise]

I'm not interested about users, I'm interested about how many of those "users" (which many are inactive anyway) participate on Steam survey. It's known fact that 1. participation is voluntary and 2. not all users are even asked to participate. You are still far from "statistically significant" amount.

Reason they refuse to tell amount of Steam survey participants is very simple. If we know 1. amount of Steam survey participants 2. how many percent have certain video card, we could easily calculate amount of those video cards participating Steam survey. And if that amount exceeds amount of manufactured video cards using that GPU, then we could immediately conclude Steam survey flawed. No wonder they don't want to take risk and refuse to tell exact numbers.

we can safely conclude Steam survey does not reflect on real GPU sales.

Even if only 5% of users complete the survey (I would wager it’s FAR higher), then we are still talking several million results… that IS statistically significant… you can cry and whine all you like, but until you get something better, just stop…

 

[HardReset]

Even if only 5% of users complete the survey (I would wager it’s FAR higher), then we are still talking several million results… that IS statistically significant… you can cry and whine all you like, but until you get something better, just stop…

If, if and if. How about finally giving some real numbers? I gave above and those without any question prove Steam survey is flawed.

 

[Squid Surprise]

If, if and if. How about finally giving some real numbers? I gave above and those without any question prove Steam survey is flawed.

I gave you numbers too… the difference was, mine were relevant - yours were not.

 

[manowar51]

You already know that is pretty much impossible to prove. However GPU sales and Steam statistics have significant difference. And that difference has to come from somewhere.

Western gaming market? From survey, Simplified Chinese and Russian alone (from top 3 only) have combined share of 36.73%. Not THAT much about western market tbh.

Intel has sold more GPUs than AMD and Nvidia combined for over a decade now. So that's not surprising at all. I'd like to see actual statistics about 5000 series sales, what actual statistics we have says that AMD sells much better. Again, if something is sold out. It may be just that amount sold was very low.

For example, Q4/2024 AMD discrete market share (AMD has pretty much displaced crap cards with APUs so AMD discrete cards are for gaming) was up around 7% and Nvidia share was down around 7%. And that was seen on Steam survey, well, it didn't. AMD just got more market share on integrated ones, in other words, AMD shipping more discrete cards means AMD got even less discrete share on Steam survey. That doesn't add up.

Here are actual sale figures: https://wccftech.com/mindfactory-week-18-2025-gpu-sales-stats-rx-9070-xt-is-top-seller/

You can say that is only one seller but those are actual units sold. Steam survey have several forged results, I know because I have made several.

Your source is poor, it just shows one small market. It doesnt give you an overall picture like the steam survey would. No intellligent person would believe that data disproves the steam survey.

By reading through your comments, you appear to have an emotional attachment to AMD. You seem upset that they have poor market share and seem to want to deny the reality of that.

Why are you so bothered that people dont buy AMD Radeon graphics cards?

 

[Squid Surprise]

Your source is poor, it just shows one small market. It doesnt give you an overall picture like the steam survey would. No intellligent person would believe that data disproves the steam survey.

By reading through your comments, you appear to have an emotional attachment to AMD. You seem upset that they have poor market share and seem to want to deny the reality of that.

Why are you so bothered that people dont buy AMD Radeon graphics cards?

If you check his other posts, he has been consistently supporting AMD for years… I’ve always assumed he is paid by them… it’s easy to agree with him when it’s about CPUs, but GPU threads generally tend to make him look ridiculous.

 

[HardReset]

I gave you numbers too… the difference was, mine were relevant - yours were not.

Relevant, but actually you don't have ANY source for that.

For some reason you missed my post regarding actual GPU shipments.

Your source is poor, it just shows one small market. It doesnt give you an overall picture like the steam survey would. No intellligent person would believe that data disproves the steam survey.

By reading through your comments, you appear to have an emotional attachment to AMD. You seem upset that they have poor market share and seem to want to deny the reality of that.

Why are you so bothered that people dont buy AMD Radeon graphics cards?

How does Steam survey give any kind of "overall picture"? There is no any source about how exactly Steam survey works. Or is there?

If you check his other posts, he has been consistently supporting AMD for years… I’ve always assumed he is paid by them… it’s easy to agree with him when it’s about CPUs, but GPU threads generally tend to make him look ridiculous.

Funny, I assume you are paid by Nvidia.

Since you also seem to miss my post where I prove I'm right, let me repost it here:

OK, here's some hard data for you:

AMD total share is: 17.51%
(Intel has: 7,78%)

Now we exclude AMD integrated solutions:

AMD Radeon Graphics 2.10%
AMD Radeon(TM) Graphics 2.01%
AMD Radeon Vega 8 Graphics 0.39%
AMD Radeon Vega 3 Graphics 0.25%
AMD Radeon (TM) Graphics 0.29%
AMD Radeon 780M Graphics 0.21%

Integrated total: 5.25%

That leaves AMD Discrete share on Steam survey to be around 12%.

But looking at AMD discrete GPU share based on shipments, that number is waaaay too low:

NVIDIA vs. AMD Discrete GPU Market Share (2010 to 2025)

Here is the NVIDIA vs. AMD discrete market share in percentage from 2010 to 2024. NVIDIA has performed extremely well in the last decade.

pcviewed.com

Also taking into account that Nvidia low end trash cards like GT710 are discrete ones (GT710 does even make it to the list anywhere) whereas AMD low end trash are integrated, we can safely conclude Steam survey does not reflect on real GPU sales.