Nvidia recommends a 750w power supply for the 3080 and 3090
AMD recommends 700w for the rx 6900xt
Do you think in the upcoming future 500-750w PSUs going to be not recommended any longer for the average build.
Perhaps starting to recommend past silver rated PSUs and start recommending gold/platinum rated?
Can’t imagine things getting too much more efficient in terms of hardware - I believe if there’s any optimization it’s going to be on the software end.
Things like DLSS/FidelityFX might pave the way to reducing hardware requirements to get more “performance” vs brute forcing it the way we do now.
Or in terms of general game engine design, like RE Engine or the Doom 2016 engine just needing less hardware to get better results.
As it stands yeah, a 750w seems to be like the bare minimum if you want a high end CPU and GPU, just to deal with the high transients these cards tend to push.
Yeah I’m fulling expecting 1000w to be the norm for more high-end systems in the next 5 years. 750w will be your basic entry point. Go back and look at power specs in the early 2000’s on cards alone, it’s just the natural progression until there is a big technology jump.
no, AMD & Nvidia seem to go in cycles coming out with power hungry chips and then refining them for better efficiency. The RTX 3080 pulls 300w gaming, pair it with a power hungry 10900k and you are at 400-425w gaming. Even if you OC you will be around 475-500w while gaming. Also efficiency ratings have nothing to do with a PSU ability to provide power, in fact they are more for marketing than anything else. Remember the testing is done at half the temp it can reach inside your PSU and the OEM gets to cherry pick the PSU they send for testing. Always go by independent testing to get an idea about actual PSU performance.
pure hyperbole unless you are at top tier parts like the 10900k & RTX 3090. You will always have people that go over their needs; typical they don’t understand PSU and how they work, they think more power = more performance, bragging right (my PSU is bigger than yours) etc.,
but that’s a small niche of people. Even prior to our current hardware situation the average gamer would only need a 550w unit. Maybe 650w if they heavily OC. Most gaming PC have a GPU that is powered by a single PCIe power connector and a CPU that may pull 100w tops while gaming.
I think the qualifiers for “high end CPU and GPU” make 750w still appropriate minimum for that group.
3080 users have reported issues with 750w units from EVGA and Seasonic resets - and that’s not even the most power hungry of cards.
For your typical mid-range users with a 60 or 70 series card, I still think 650w is a better recommendation than the 500w of old to overcome some of the inadequacies of lower rated supplies actually reaching their stated ratings.
Not argument here, I think a quality 750w handles “most” upper tier PC parts. Obviously as you increase extreme power needs you need a high power PSU.
I’ve heard this but I’ve yet to encounter someone I know and trust to post accurate info on the issue. That said, not necessarily surprised it’s from those OEMS (SF for EVGA) and not guys like CWT or Delta. It may not necessarily be a power demand issue as much as the GPU demands are going out of spec.
The old adage of higher power rating doesn’t make a PSU better, better parts and better build make a PSU better.
I haven’t experienced it either myself, and supposedly Nvidia has fixed a lot of the transient problems with their later drivers.
Most of the people who have experienced it is only lower rated versions of my PSU. Even if they are in-spec, the higher rated units have higher thresholds before tripping.
Keep in mind this unit was very well reviewed by Johnny Guru and other forums prior to the discovery of the transient load issues and tripping.
Still, even if we write these off as the fault of the graphics card manufacturers, having a more capable supply in real life ultimately resulted in a better experience - users of 550w and 650w tripped, and users with 850w and 1000w didn’t - so either way if it is the fault of the graphics card or not - getting “more” was just better.
Now, that doesn’t mean I think we should be going out to grab 1000w units from some no-name manufacturer just for that reason, but I think in a practical world getting more than you need is still not a terrible idea, given these things seem to pop up so often.
So instead of thinking, "I should get more than I need in case I want more powerful graphics cards, I think, “what if my parts aren’t in spec and need more than they actually ask for?”
I believe I have the same family unit as you or at least they share a similar platform & parts. As good of a review that guys like Jeremy and Aris do, OEMs also focus on getting those epic stability and efficiency numbers to get the raving reviews. Yet the smaller guys tend to skip on QC testing (because of cost in machinery & labor) and simply focus on building a PSU to get reviews. If the brand wants a full run of QC, they have to pay for it out of house since the smaller OEM can’t provide it. There is a big difference with a large OEM with its own SMT lines and full QC capability and smaller OEM who’s SMT is done third party (or off site) and relies on the brand for full QC testing. Even Aris has stated he won’t test PSU at 50c (rather 47c) because he knows a number of the “quality” units won’t pass the test. Just because a manufacturer is “tier 1 
” in reviews does not mean they are a “tier 1” manufacturer in capability.
When I look for a PSU, I do the following
Does it have the connectors I need
Does it provide the performance I want (as proven by a 3rd party)
If I believe a 650w unit can’t handle a PC that games at 325w and may spike to 400w or that I do an upgrade that adds 50w to 75w to my current hardware and the PSU can’t handle it than I am looking at the wrong PSU and I move on.
The lack of good testing further reinforces my belief that buying bigger (on the ones we assume are good based on the available reviews, with all the caveats those bring) makes sense.
Buying a bit more than I needed this last time worked out in my favor, so even if it’s extraneous it will probably be the way I go moving forward 
Plenty of good testing done by flextronics, delta, cwt, fsp among others. It’s why when certain brands go with a specific OEM, cap manufacturer, etc., over another and get crap for it, I know the engineers are rolling their eyes.
Personally speaking for a long time now my “minimum” PSU for systems with a non-trivial GPU has been 650W by a known good OEM. In my mind PSUs aren’t something to screw around with, so it feels safer to have a decent bit of margin, both to ensure that normal usage is absolutely never going to cause problems and to ensure there’s headroom for upgrades. It’s served me well so far, in over a decade of building PCs none have ever had power delivery issues.
But I also don’t usually build with top-end parts. The closest I’ve ever been was with my 6700k + 980Ti build, and neither of those are particularly power hungry. Since then purchases have been more mid-tier.
I’ve been running at least a 750w for an insanely long time, my last computer ran 11 years on the same one with zero issue. Keep in mind when it comes to electrical anything I’d rather have more than I need all the time. Its not much different than a car engine, if you can run it at 75 at 2000 rpms you would much rather do that than running it at 4000. It works fine but its just asking for it imo.
not the same and yes especially if you are charging out of a corner in a race car…or worse on a bike
I’m talking more interstate driving. I mean race cars are built to run high anyways.
The reality is your fan is typically the first thing to go in most PSU (assuming they haven’t cut other major corners) so its quality and how long it gets used will have more to do with PSU life span for most people. Of course you have people going full load on a cheap PSU but that’s like buying a yugo and then complaining it broke after 10k miles.
Well I’ll stick with what has worked for me. Been buying computers since the late 80’s and never had a PSU fail. All of them have gotten heavy work and a ton of hours. Why deal with issues when there is no reason too.
Unless the PC in question is constantly running at full tilt, wouldn’t buying PSUs designed to run fanless most of the time (e.g. at idle and below a certain threshold) almost guarantee that their fans won’t wear out? Unless vendors intentionally use cheaper fans in such models…