News Nvidia RTX 4060 Arrives on Steam Survey, Matching RX 7000 Market Share

[Admin]

Three months after launch, Nvidia's mid-range RTX 4060 makes an appearance in the Steam Hardware Survey for August. It debuts with an equal percentage to AMD's RX 7900 XTX, the only RDNA 3 GPU currently on the list.

Nvidia RTX 4060 Arrives on Steam Survey, Matching RX 7000 Market Share : Read more

 

[NeoMorpheus]

These figures really show how much more popular Nvidia's GPUs are, no matter how competitive AMD's GPUs may be, or how poorly the cards might be received by reviewers and influencers.

Actually, the influencers and biased "reviewers" have more pull than real reviewers.

Those are the ones that will say "the 4060 is not a good gpu, but the magical DLSS and fake frames will make it fly so go ahead a buy!"

The same will also say "AMD is DOA, regardless if its faster or cheaper, buy Nvidia Yo!!".

Yeah, we are doomed.

That said, I think that the only way we can ever get proper results is if Valve does a survey to 100% endpoints, not just some random systems. But it is the closest thing we have to an accurate count.

 

[PEnns]

Actually, the influencers and biased "reviewers" have more pull than real reviewers.

Those are the ones that will say "the 4060 is not a good gpu, but the magical DLSS and fake frames will make it fly so go ahead a buy!"

Not to mention the usual websites.....

 

[USAFRet]

Lets leave the fanboyism out of this.

Instead, speak to the actual products.

 

[CelicaGT]

Comparing 40 series adoption to 30 series and saying it's better? 30 series availability for gamers was abysmal (scalpers, miners), 40 series are readily available. These numbers do not mean what the article implies they mean given the timeline....

 

[Elusive Ruse]

Evidence that the majority of dGPU purchases are made ignorant of competition.

 

[thisisaname]

Actually, the influencers and biased "reviewers" have more pull than real reviewers.

Those are the ones that will say "the 4060 is not a good gpu, but the magical DLSS and fake frames will make it fly so go ahead a buy!"

The same will also say "AMD is DOA, regardless if its faster or cheaper, buy Nvidia Yo!!".

Yeah, we are doomed.

That said, I think that the only way we can ever get proper results is if Valve does a survey to 100% endpoints, not just some random systems. But it is the closest thing we have to an accurate count.

Someone wiser than me once said "There is not such thing as a bad gpu just a bad price"

In this case the 4060 is far to cut down for it price.

 

[JarredWaltonGPU]

Comparing 40 series adoption to 30 series and saying it's better? 30 series availability for gamers was abysmal (scalpers, miners), 40 series are readily available. These numbers do not mean what the article implies they mean given the timeline....

We shall see in the coming months. I would bet heavily that RTX 4060 will do very well on the survey over time. Because it is actually faster than the 3060 while costing less. That’s a potent combination.

And if you think the 8GB will stop adoption, just look at how many people bought the 2060. It might not pass the 3060 for a while, but it’s virtually guaranteed to hit the top five or so GPUs over the next year or two.

And if the street price drops closer to $250 over time, it will do even better.

 

[NaviiX]

We shall see in the coming months. I would bet heavily that RTX 4060 will do very well on the survey over time. Because it is actually faster than the 3060 while costing less. That’s a potent combination.

And if you think the 8GB will stop adoption, just look at how many people bought the 2060. It might not pass the 3060 for a while, but it’s virtually guaranteed to hit the top five or so GPUs over the next year or two.

And if the street price drops closer to $250 over time, it will do even better.

100 percent agree. Radeon isn't competing at the low end or the high end with RX 7000. You only buy Radeon if you care about pure raster. The $500 7800xt is decent but lacking features that people care about. The big bump in raster for $50 vs 4060 Ti 16gb or the $100 saved for a 4070 won't convince most people.

AMD needs to win mind share and compete at the low end and midrange much harder, even if it's at a loss. They need to invest into deep learning to make FSR look better and push FSR3 harder. Now Nvidia is coming with Ray reconstruction which looks pretty incredible. It's not 2020 anymore and DLSS/RT seriously matter, that's ignoring things like reflex, ray reconstruction, or frame gen.

Not to hype anyone to rush to buy GPU's. If your GPU works fine in the games you play keep it for now. Radeon is just failing miserably with their product stack, in my opinion. Nvidia fumbled the ball directly into their hands and they're fumbling it right back.

 

[Colif]

So it doesn't really help that it doesn't automatically run the survey after you update hardware

Unfortunately, what makes Steam Hardware Survey kick off is about as mysterious as confirmation of Half Life 3. If you swapped out your RAM, motherboard, a new higher resolution monitor, or even a new CPU, you may or may not have the survey kick off.

But you can run it manually if you want to:

First, press the Windows Key + the R key simultaneously to summon Windows 10’s “Run” prompt. Now, enter one of the four lines:

steam://takesurvey/1/
steam://takesurvey/2/
steam://takesurvey/3/
steam://takesurvey/4/

How to make your PC take the Steam Hardware Survey

Make your PC hardware count by taking the Steam Hardware Survey.

www.pcworld.com

So its accuracy is questionable unless they can make it run every year or so to confirm figures.

made me wonder as I played Steam games all year on a 7900XT so surprised none are showing at all.

 

[Elusive Ruse]

100 percent agree. Radeon isn't competing at the low end or the high end with RX 7000. You only buy Radeon if you care about pure raster. The $500 7800xt is decent but lacking features that people care about. The big bump in raster for $50 vs 4060 Ti 16gb or the $100 saved for a 4070 won't convince most people.

AMD needs to win mind share and compete at the low end and midrange much harder, even if it's at a loss. They need to invest into deep learning to make FSR look better and push FSR3 harder. Now Nvidia is coming with Ray reconstruction which looks pretty incredible. It's not 2020 anymore and DLSS/RT seriously matter, that's ignoring things like reflex, ray reconstruction, or frame gen.

Not to hype anyone to rush to buy GPU's. If your GPU works fine in the games you play keep it for now. Radeon is just failing miserably with their product stack, in my opinion. Nvidia fumbled the ball directly into their hands and they're fumbling it right back.

What is there not to get convinced unless one is a mindless Nvidia buyer? The 7800 XT beats the 4070 in raster and trumps 4060Ti while also beating it in RT, it costs as much as the 4060Ti and $100 less than the 4070. Pure raster wins over RT performance any day, RayTracing is icing on the cake not the whole cake. Your post reeks of misdirection and misinformation which makes me question your motives.

 

[NaviiX]

What is there not to get convinced unless one is a mindless Nvidia buyer? The 7800 XT beats the 4070 in raster and trumps 4060Ti while also beating it in RT, it costs as much as the 4060Ti and $100 less than the 4070. Pure raster wins over RT performance any day, RayTracing is icing on the cake not the whole cake. Your post reeks of misdirection and misinformation which makes me question your motives.

You underestimate how many people are willing to either spend the extra 100 dollars for Nvidia GPU features and/or lose raster performance. Personally a lot of the games i play utilize DLSS/DLSS3 and I think FSR looks horrific so I wouldn't buy an AMD GPU. Good luck convincing the masses by calling them brainless, surely that will work.

You don't realize it because you're brainless but you are doing for AMD what you are accusing me of doing. People should make their own decisions. I hate Nvidia's marketing but there is just some facts that AMD fanboys continue to cope with. Otherwise, why is AMD scrambling to drop prices and catch up with FSR3 / FSR / Antilag + ? If they don't drop prices further and catch up in software they will not eat market share.

 

[USAFRet]

To all involved here - No more personal attacks.
This is not negotiable.

 

[Elusive Ruse]

You underestimate how many people are willing to either spend the extra 100 dollars for Nvidia GPU features and/or lose raster performance. Personally a lot of the games i play utilize DLSS/DLSS3 and I think FSR looks horrific so I wouldn't buy an AMD GPU. Good luck convincing the masses by calling them brainless, surely that will work.

You don't realize it because you're brainless but you are doing for AMD what you are accusing me of doing. People should make their own decisions. I hate Nvidia's marketing but there is just some facts that AMD fanboys continue to cope with. Otherwise, why is AMD scrambling to drop prices and catch up with FSR3 / FSR / Antilag + ? If they don't drop prices further and catch up in software they will not eat market share.

So all you got as an argument is the masses are willing to pay $100 more just to get Nvidia, well yea, that's the whole point. Even you admit that there's no logic behind it. I got an Nvidia GPU in my PC because when it came out it was the best choice for my budget. Next time I buy a GPU which is going to be soon I'm most probably going for an AMD GPU cause now that's where the best bang for the buck is.
That is the difference between you and I. I buy based on what gives me the best price to performance, you buy because you want to have Nvidia.

 

[Unolocogringo]

What most people fail to take into consideration is that Nvidia has +/- 80% of GPU market share.
There are 4 Nvidia cards sold for every 1 AMD card sold.
Most of those being entry/mid range cards..
So Steam surveys will always have a higher share of Nvidia cards.

 

[NeoMorpheus]

So it doesn't really help that it doesn't automatically run the survey after you update hardware



But you can run it manually if you want to:

How to make your PC take the Steam Hardware Survey

Make your PC hardware count by taking the Steam Hardware Survey.

www.pcworld.com

So its accuracy is questionable unless they can make it run every year or so to confirm figures.

made me wonder as I played Steam games all year on a 7900XT so surprised none are showing at all.

A couple of points to your post:

1- As i said above, the accuracy is very questionable and can only be acceptable if everyone was counted, not just some random systems.

2- Expanding on point 1, many people (myself included) reports during many years to rarely receive a survey when playing on a system with an AMD gpu, but will get one every month when using a ngreedia one. As you said, nobody seems to know why.

3- I could be wrong, but I read that manually triggering the survey does not get that system counted, hence doesnt affect the overall numbers in the survey.

Personally a lot of the games i play utilize DLSS/DLSS3 and I think FSR looks horrific so I wouldn't buy an AMD GPU.

This is why i have a problem when reviewers/influencers keep pushing and worshiping dlss and never, ever call out this very action as anticonsumer and a lock-in tech that its taking the openness of pc gaming away.

Its what they have tried over and over with tech like PhysX, gameworks and now this.

Many games are permanently locked into their hardware, like the Arkham games, which will only show all the graphics details if you are playing with their hardware. All other platforms can pound sand and many gamers are not even aware of what they are missing.

How can these reviewers/influencers that are not biased or bribed keep ignoring that in good faith? I simply dont know, but the damage its done, as your comment confirms.

 

[Colif]

2- Expanding on point 1, many people (myself included) reports during many years to rarely receive a survey when playing on a system with an AMD gpu, but will get one every month when using a ngreedia one. As you said, nobody seems to know why.

the 4 surveys I showed seem to match to specific games/activities

The only difference between the surveys appears to be what game or app the the survey is tied to, but all seem to collect identical information. If you enter 1, you’re taking the Dota 2 survey; 2 is for Counter Strike: Global Offensive; 3 is for Source Filmmaker; and 4 is the Steam VR survey. You can only take each one once.

And I don't have any idea if there are more games tied to surveys.
I have never seen it but I tried one of the lines and all that happened was steam opened. It didn't show any data collection as such. Steam does appear to know what GPU I have now.

TBH I don't really take any notice of surveys like this when its obvious most of the people have cards from one maker.

 

[NeoMorpheus]

the 4 surveys I showed seem to match to specific games/activities

And I don't have any idea if there are more games tied to surveys.
I have never seen it but I tried one of the lines and all that happened was steam opened. It didn't show any data collection as such. Steam does appear to know what GPU I have now.

TBH I don't really take any notice of surveys like this when its obvious most of the people have cards from one maker.

Hmm, interesting.

In my case, the very few times that I received a survey (in 10 years of usage), was right after launching Steam itself, not by starting a game.

 

[KyaraM]

100 percent agree. Radeon isn't competing at the low end or the high end with RX 7000. You only buy Radeon if you care about pure raster. The $500 7800xt is decent but lacking features that people care about. The big bump in raster for $50 vs 4060 Ti 16gb or the $100 saved for a 4070 won't convince most people.

AMD needs to win mind share and compete at the low end and midrange much harder, even if it's at a loss. They need to invest into deep learning to make FSR look better and push FSR3 harder. Now Nvidia is coming with Ray reconstruction which looks pretty incredible. It's not 2020 anymore and DLSS/RT seriously matter, that's ignoring things like reflex, ray reconstruction, or frame gen.

Not to hype anyone to rush to buy GPU's. If your GPU works fine in the games you play keep it for now. Radeon is just failing miserably with their product stack, in my opinion. Nvidia fumbled the ball directly into their hands and they're fumbling it right back.

Plus, Steam survey doesn't just cover the US, and prices can vary widely per country. For example, to pick your own comparison, at time of writing the 4060Ti 16GB costs 462€, the 7800XT 559€ and the 4070 costs 589€ in Germany, and yesterday I have seen two 4070s costing only 549€. Much more sensible price deltas relative to performance and feature set I think (though the 4060Ti still needs to drop lower) and essentially the reverse price situation compared to the US. Heck, not even the 7700XT is all that much better positioned at 485€, though that one offers rather good performance all around for the price especially compared to the 4060Ti. This is also an issue for AMD here; performance isn't much larger than the 4070, but RT performance much lower and the Nvidia feature set is better, so why get the Radeon at the current time when the 4070 costs only 30€ more, or the same depending on model since there are huge price overlaps? Plus, Nvidia pushed out two pretty great features this year, while AMD struggles to catch up with the arguably less important/more divisive one of them with no word about the other yet. Not a good showing from their part, no matter how you look at it.

That naturally influences buyers and how popular certain GPUs are as well, if GPU X is cheaper than GPU Y in enough countries, and only your own country has the deltas etc differently, then it doesn't really matter how the situation is in your country; the survey will reflect the international market. Also, the prebuilt market plays a huge role as well and cannot be disregarded.

I fully agree with what you say there, too. AMD had a real chance to gain market share this year via pricing, and they completely botched it. Literally every single one of their cards would have to be priced much lower at launch to attract people, or offer better RT performance and features; especially the latter simply cannot compete. At all. People can moan about proprietary features all they like, if there is no alternative to them, they win handsdown. Most people only care if thoae features exist and work, not if they are proprietary or not.

 

[NeoMorpheus]

offer better RT performance

I honestly wished that people would stop repeating that just because reviewers do.

As said before RT its simply useless (so far) given the crazy performance hit that doesnt translate to anything that justify such hit.

There are a couple of games out of thousands of games available that show a bit of eye candy due to RT, but really, in its current form and foreseeable future, RT should not get that much "demand" that everyone keeps repeating.

 

[JarredWaltonGPU]

So it doesn't really help that it doesn't automatically run the survey after you update hardware

But you can run it manually if you want to:

....

So its accuracy is questionable unless they can make it run every year or so to confirm figures.

made me wonder as I played Steam games all year on a 7900XT so surprised none are showing at all.

Proper statistics is that the survey should run randomly, with no way to circumvent that. It doesn't matter that you just upgraded hardware! You're still just a single potential sample point. If it ran automatically for everyone that swapped a component, it would skew massively toward new hardware.

The real problem is that Valve has never given any fully clear indication of how it does the sampling. Is it truly random, like it should be? We don't know. Probably. But then you were able to give instructions that show how to circumvent the random sampling. How many people are using that to mess up the data gathering, and can Valve do anything to eliminate those results? Again, we don't know.

What Valve should do is provide some formalized statement on the statistics used. Something like: "We use random sampling, gathering data from over 100,000 selected Steam PCs every month. This gives us a 95% confidence level and helps to explain minor month-to-month fluctuations."

It would also be good to automatically pull out any GPU with over 0.01% of the market share. It looks like Valve does this on the API page, but I seriously doubt that's correct. Otherwise, why are there still 0.73% of GPUs that are "unknown," and are you really telling me that the 7900 XTX has 0.23% (but doesn't show up on the API pages!), while the 7900 XT and 7600 have less than 0.01%? I call shenanigans!

More critically, why do all of the numbers for DirectX 12 sum up to 89.92% in August? That literally makes zero sense. I could see if the numbers summed to maybe 98% or more, but there are only 213 GPUs for DX12. With a maximum 0.01 rounding error, that means only a 2.13% total error would be possible. That would require every GPU to be rounded down from for example 0.0199 to 0.01, and the total would still sum up to around 98%. Which means Valve is doing something else, without explanation. (And the Vulkan numbers are even more stupid, as they sum up to around 180% total — looks like everything gets counted twice.)