Trends from the April 2026 Steam Hardware Survey

DAPUNISHER said:
MSI? If so, it would track with why they bailed on radeon.

LTT had tens of thousands more responses than the TPU poll, same roughly 1 in 3 users.
Click to expand...
I would say the population of LTT (or TPU) readers is not a population that is representative of all steam users...
 
DAPUNISHER said:
They already have the same basic system specs telemetry e.g. AMD collects it via radeon software. It is all they need to sample. It contains OS, gpu, cpu, amount of ram.
Click to expand...
I'm thinking more along the lines of enough telemetry to uniquely identify a system and an install - like, if you wipe your OS between surveys - or hell, just reboot into another OS (dual-boot Windows and Linux) - that would influence the survey results. The question is whether Valve collects enough data to make that distinction, and whether the FOSS / privacy communities would be comfortable if they did.

I'm betting that there's something falling through the cracks here that requires some cleanup to interpret.

DAPUNISHER said:
I have all 3 vendors cards, your anecdote is faulty.
Click to expand...
LazyGamer said:
and that comes with side effects that are likely more vocal and AMD users...
Click to expand...
This was supposed to be "than AMD users".
 
So I assume there is no published methodology, so its complete garbage.
Shoulnt be discussed at all, seriously.
It may be a long running experiment /PR op by some valve employee for all we know, to see how info spreads, reactions and such matters, not saying malicious, plus bonus keeps you talked about, for free.

As mentioned, I too cant see the difficulty of having hyper accurate data on, gpu, cpu, ram, hdd and a number of hardware aspecs are core part of programs, more so games funcioning correctly so why would the steam client itself not have data.
Are we forgetting each and every chip since the 90s? Has a unique id number, this across basically every cpu ever made arm x86 whatever.
The client likely has everything, everysingle cpu id everything. No names and adresses needed.
Fantasy to think otherwise.
Why they dont use it, idk, probably its trade secret style data, hence do this weird PR survey crap. Matter of fact even if they mention methodology, i convinced myself its a PR tool end of story.
 
David_Schroth said:
I would say the population of LTT (or TPU) readers is not a population that is representative of all steam users...
Click to expand...
Agreed. Steam is a huge platform, no doubt. It does seem highly likely the polls are mostly limited to/pertain to non Asian gamers. However, LTT has almost 17M subs on their main channel alone. And both polls had far beyond the number of responses necessary for accuracy. The results reflect millions in those regions, and is worth inclusion, despite not representing all steam users everywhere.

Of course, AMD graphics potentially being under represented does not address the problems with the Steam survey itself. I also do not subscribe to the hypothesis that there is a conspiracy against AMD. Valve has worked closely with AMD and used their hardware for both the Steam Deck and Steam Machine. My position on the topic is that the Steam survey should be taken with a grain of salt. Given it is a black box prone to enormous errors.
 
Uvilla said:
Why they dont use it, idk, probably its trade secret style data, hence do this weird PR survey crap. Matter of fact even if they mention methodology, i convinced myself its a PR tool end of story.
Click to expand...
Interesting take on it. I had not considered it from a solely PR/marketing angle.

AMD cited the telemetry from the Adrenalin app when discussing the real possibility of manufacturing some SKUs again. It is a straight forward use of the data collected anonymously. It makes your idea that the survey is a PR gimmick, rather appealing to me as an explanation for why it is even conducted.
 
GodisanAtheist said:
I know it's a bit of a conspiracy at this point, but I don't think AMD cards get surveyed as often as Nvidia cards.
Click to expand...
Conspiracy are often real, and proven correct. This is a prime example where that is the case. It is more broken than simply not being surveyed though. https://www.techpowerup.com/349940/...indows-11-adoption-amd-gpu-detection-fix-live

At almost 20% share AMD is still under represented in the survey. The survey has always been nerfed and likely always will be. They should just use the telemetry and stop this supposedly random sampling opt in nonsense which is constantly bugged and broken. Hanlon's razor.
LazyGamer said:
However, I would argue that the relatively small marketshare of AMD GPUs might have something to do with it; their users are typically louder than than their balance sheets support...
Click to expand...
Turns out where there was smoke, there was fire. it was indeed the survey being nerfed. JP Research also ended up revising their estimate for AMD GPU shippment numbers in Q4 upwards by 400000 units. That's a big goof too.
 
Brian_B said:
Really? There are a metric $***ton of bat$*** theories out there.

I might accept “occasionally” - but “often”?
Click to expand...
I concede the point. Now, lets get back to the conversations concerning the fact the survey once again has proven to be broken. And this is only the latest incident proving it.
 
DAPUNISHER said:
Turns out where there was smoke, there was fire. it was indeed the survey being nerfed. JP Research also ended up revising their estimate for AMD GPU shippment numbers in Q4 upwards by 400000 units. That's a big goof too.
Click to expand...

It seems so. They went from 5% Steam marketshare to 8%... which is a lot, but doesn't really move the needle into the same bracket as their shipments?
 
LazyGamer said:
It seems so. They went from 5% Steam marketshare to 8%... which is a lot, but doesn't really move the needle into the same bracket as their shipments?
Click to expand...
Exactly. And we have no idea if it is remotely accurate. What we do know is AMD has reported strong sales YOY the last EDIT: 3 Qs.
"Gaming revenue was $1.3 billion, up 181% year-over-year driven by higher semi-custom revenue and strong demand for Radeon™ gaming GPUs."
Click to expand...
From Q4 last year
"Gaming business revenue in the quarter was $843 million, up 50% year-over-year, primarily driven by higher semi-custom revenue and strong demand for AMD Radeon™ GPUs."
Click to expand...
And Q1
Client and Gaming segment revenue was $3.6 billion, up 23% year-over-year. Client business revenue was $2.9 billion, up 26% year-over-year, primarily driven by strong demand for leadership AMD Ryzen processors and continued market share gains. Gaming business revenue was $720 million, up 11% year-over-year, driven by solid demand for AMD Radeon GPUs partially offset by lower semi-custom revenue.
Click to expand...
My contention is; The Steam survey is too often a case of bad data is worse than no data. The old truism about lies, damned lies, and statistics. It ends up being misleading. Hence, it is best to take it with a grain of salt. That's it, that's all I am suggesting.
 
Last edited:
LazyGamer said:
I find it frustrating that the data, which as a sample is far beyond what is needed for statistical representation, would be so poorly collected.
Click to expand...
Exactly. Just used the damned telemetry they get as part of the TOS. While they are at it, they can break it down by regions. For example; I may be confidently incorrect, but I expect Nvidia is far more dominant in Asia than the Americas and Europe.

The idea that it is a PR tool makes a lot of sense. News cycle engagement bait.
 
You must log in or register to reply here.
Become a Patron!
Back
Top