im just ignoring this gen because its so absurdly priced. $1000 USD for an 80 class card??? screw that, ill wait for something actually good thx

It seems like my salary doesn't even bother catching up to inflation

RX 6600 has to be everyone’s go to entry-level GPU at this point. In some countries you could find them used for under $120 and there’s virtually no competition at that price point

exactly my situation. sitting tight on my 1070. Couldn't get a hold of any GPU during the crypto boom, so waited out for 40 series. What a disappointment.

I upgrade every 7 years unless the card itself commits seppuku

I went from buying a card every gen to making them last as long as possible. In other words, I used to upgrade because I loved the new hardware. Now I only upgrade out of necessity.

Another important thing to note is that while prices rose thanks to inflation, most people's wages didn't. So many people can't really afford to shop at the same performance tier anymore.

Oh absolutely do a follow up on the High End too. Also "predict" what the next Gen "should or could be" if you take those Numbers.

I will not bend over and pay these prices

Nostalgically looking back at an age when a mid xx70 card was just £300.

Companies like nvidia had two possible paths to follow: Selling lots of video cards at a low price, or selling less video cards at a premium price and obtain the same net profit... obviously, they've chosen the latter. I've been building my own Pcs since early 2000s and I always upgraded my video card every christmas. The performance upgrade was always totally worth it. Now I'm just stuck with my 9th gen cpu and my 3060ti, why? because current prices feel like they're scamming you, I can easily pay for the greatest hardware out there, but I cant stop thinking than half of what I'm paying is going directly to some CEO's pocket who randomly put a price tag on a product that should be cheaper...

My biggest struggle with Nvidia is that they lock their cards in every way possible (BIOS, Voltage, Power Level, PCI-E lanes, ect.) and only put the minimum amount of reasonable VRAM in. It is like paying extra to make it worse. Like buying a Lambo with 700hp but it is restricted to 100mph and the restriction is obligated and costs extra.

The biggest Problem of the 3060 and the 4060 are, that they're RTX cards, although not capable of delivering good raytracing performance. And so they're completely overpriced, while promising a feature they can not properly provide.

In a lot of cases it is even worse than that because local retailers (in Greece for example) add another 50/100 euros, or more, to already overpriced products, making them even less worthwhile.

Great video Tim! I upgraded from a GTX 970 to a GTX 1080 for about $130US this year. Nothing seems to be worth it for 3x the cost or more.

xx60 8GB - $250, xx60ti 12GB - $300 xx70 12GB - $400, xx70ti 16GB - $500 xx80 16GB - $700, xx80ti 20GB - $800 xx90 with 24GB+ - $1000 (but as a HALO product for semi-professional use for AI it could be $1200)

Really unique video proposition. We need to keep doing this every release. Good job!

6:00 "But surprisingly, they've actually kept pace, offering the same level of terribleness each time." 👈👈 That sums up Nvidia pretty nicely. And why I haven't upgraded since 2016.

I upgrade every 4-5 Years~. My goals are not based on % of the GPU performance jumps per generation. My expectations are linked to FPS. I do not upgrade unless i get under 60 FPS regularly on different games. Example: If i can play 4 out of 5 games (tripple A) at 60 FPS or higher i see no reason to upgrade. At 2 out of 5 games under 60 FPS...its time to think about upgrades. Currently running a 6900 XT Nitro and a R5 5600x for 3 Years now and they are both still going super strong in 1440p. Might swap with RX 8000 and new R5 8000 CPUS.

I miss the time when 200$ buys you a decent GPU, and budget segment was 150$ or below.

When I upgrade my GPU I just want something that is twice as good as what I've got, and it's the same with my CPU upgrades as well.

Title gets a thumbs up from that alone. Yes the prices do indeed suck.

This was one awesome analysis, well done! I only miss the same repricing done for the highest-end GPUs, so yes definitely that followup would be much welcome!

man, Thank you for this video, its great to see validation in our minds to see how things haven't really kept up!

Laughing while seeing the 7900XT costing 1000USD and the 7900 XTX costing 1200USD, here in Brazil, even after AMD cut prices. Nvidia is even more expensive, im not buying any graphics card on these stupid prices anytime soon.

makes sense the market cap of nvidia is higher than the gdp of canada.

Loved the video and especially the in depth analysis and inclusion of so many graphics cards. Can't get content like this anywhere else on YT!

The way you show (and thus sum up) the problem on these charts is great. Thank you!

I would love a part 2 for the high end

I upgraded to 6800 XT from 1660 super after around 3 years and i could not be happier as the performance leap for the price paid is so worth it.

Love this video and data comparison. Would love to see the higher side of the market. Love your guy's channel. Keep up the great work.

Hi Tim, awesome analysis, it really gave a great visualization to a feeling of lackluster products and too high prices. Id really love a high end analysis too since there expectations are even higher and small percentages can result in quite a lot of money. And you already showed that the low to mid tier products are about 20% too expensive.

Like said back in scalper pricing days, people should not buy overpriced gpus. I didnt. Now we all suffer, because people just had to have them at any price.

I'm starting to think that Nvidia and AMD are not really in competition and more so price fixing. The head of each company are actually family members. These prices are just out of control. The Ultra budget list seems to be the only one close to realistic pricing, the rest just go into the stratosphere with the prices.

It is great to see that you tackle such a complex data-driven content. I think you did a great job in presenting the data without letting the amount of numbers overwhelm the viewer.

I feel like the people who upgrade less frequently will care more about VRAM, so even if you're on a 10 series card, the regular 4060ti and below are dead in the water.

It's not about a set number of generations before upgrading. I have a GTX 1080Ti that I came much closer to upgrading to the 3080 when it came out, than anything after that. I don't like to feel fleeced, and at 1440p the GTX1080ti is still a monster. I won't upgrade until I get a 600 USD card (what I paid for the GTX1080Ti) that doubles the framerate, which isn't much to ask when you're upgrading from a 2017 card, no?

With all these booms I simply forgot what a new good offering looks like.

Amazing information, thanks for sharing this data and I would love to see other videos like this one!

Awesome comparison, love the vid!

Pretty cool research and graph work! Love to see it.

1.5 years ago i found a small shop that was going bankrupt and he was selling most of his stuff for literally whatever prices i got a 6700xt for 280 euros a 5700x for 210 32 gig ram for 68 euros

Great Video, took me a bit to get the graphs but once I did it was really informative and entertaining - Definitely would love to see the higher end and a look at the coming generations!!

Excellent presentation! Love the data eccentric approach! Awesome work!!!

From the perspective of a statistician (me), the way this data has been interpreted isn't quite correct. You are measuring Upgrade Cycle (Independent Variable), Relative Performance Uplift (Mediator/Moderator), and Satisfaction (Dependent Variable). You need to keep this in mind. Your Demographics look as follows using cumulative percentages for 80% of Population Satisfied: Every Generation: 150% (because this is how you scaled your later graphs) Two Generations: 180% Three Generation: 200% Four Generations: 200% Five Generations: 300% You only have 3 upgrade cycles mapped to your graphs over 4 years, so we can't test Four Gen and Five Gen (Although Four is the same as three so we can say a little about that). The constant shifting of the trend line to new baseline positions is interesting, but the way the interpretation continues to roll along with the shift isn't quite interpreting this data correctly. You need to interpret the effect of Upgrade Cycle choice on Satisfaction with this type of data (also use cumulative % for your graph at 1:40 - it took me two minutes to convert it). The mechanism driving this is Relative Performance Uplift - but remember that this is not your dependent variable so you shouldn't be interpreting it like a dependent variable (it is probably a moderator). Across all of the graphs, products released immediately after the previous one have a lower chance of meeting the requirements for "satisfaction" year-on-year (because this is 150%) and meeting this requirement on a rolling basis will fall short for all low performing years. This doesn't mean that these cards are a "bad buy", because that is not what you are testing here. It means that people who buy every generation are less likely to be satisfied with the Performance Uplift, compared to those who upgrade every three generations or four generations (who both require 200%). Even two year upgraders are more likely to be satisfied (at their 180%). This is because graphics card performance uplift is always going to be unstable year on year. There will be some bad years and some good years, as you have pointed out. You should resist the urge to use this data to conflate "bad buy" with a particular generation (if you are indeed basing this on satisfaction rates). If "good buy" is linked with satisfaction, then the lag between purchases is the thing to talk about. You should recommend that to avoid this "good year/bad year problem", people upgrade every 3 or 4 years, which you definitely do start alluding to at 19:25! You are overstretching to suggest that the data is telling GPU companies to change their approach (as only 9% of buyers are year-on-year buyers). Two-year, Three-year, and Four-year buyers are 75% of the market (at 0:59)! Your graphs suggest that these buyers are very likely to meet their thresholds for satisfaction (at 80% of population thresholds). More respect should be shown for your 4-year timeframe - use it! This is a good move you have made! Many fields use 5-year cycles to create rolling norms for performance data, because year-on-year is often invalid, and you can see here why. So you are following an approach approximating good practice here, but you should interpret this as testing the relationship between Upgrade Cycle (Independent Var), Relative Performance Uplift (Mediator/Moderator), and Satisfaction (Dependent Variable) - not GPU performance. Whether a GPU is a good buy (in the absolute sense), relative to other factors is highly context sensitive, and your normal approach to reviews are a far superior means of making such a judgment. Feel free to reach out to me if you want to work on your methodology.

One of your best videos! good work

Great video and a lot of good information here. This concept was well thought out and helpful. I'm down to see you do the same graphs for the high end/enthusiast cards as well.

nice analysis! I appreciate the work HUB! I agree, though their reasoning of pricier production cost would be the counter from this ideal analysis me as a consumer facing a market of duopoly business, operating costs are something they need to sweep first before making any retail steps

Ngreedias next gen gpus will be made on the disappointwell architecture.

Please do a prediction of what the 50 and 8000 series cards need to offer, that would be super interesting to compare when they actually hit shelves

What a great video. Really well done on the math and explanation side. I need to make this type of analysis often for my engineering work and i am pretty sure i am going to use this format in future.

Great work, you guys should include this type of graph when reviewing a new release

me looking at gtx 1050 gtx 1050ti and gtx 1650 and 1650 low profile second hand all at the same price, and higher then it was sold new...love Australia...

I needed an upgrade but wasn't happy with current gen prices, so I just bought used last gen. Best decision for me. Saved money and happy with the performance

This video was awesome! I would very much enjoy a high end video in this same format!

Thanks Tim, great info, please do the higher end models next.

Please make a next gen pricing video based on these calculations 🙏

This current pricing of the GPU market is driving away from PC gaming and actually spend more time with family, friends and personal upskilling for work which is a good thing actually. I don't want to feel stressed and spend $1000 for a GPU that probably won't last 3 years if games are going to have really bad optimizations. The world has gone sideways and I guess I will take a step back and think what's best for myself. I hope PC gamers will too and hopefully that will force them to lower the prices down until the old norm again.

This is an an amazing video. Brilliant work sifting through all the data to get this information. Great journalism guys! Keep up the good work!!! 🔥🔥🔥

Great video! Thx Tim! The visualization was a bit confusing and it felt rushed. I would like to ask for more videos like this, with more visualization, more time spent explaining the charts and less gpus focused on.

Is it safe to upgrade from the 1080Ti yet?

Got my Vega 56 for $240 in 2019, I've yet to find a better price/performance card with more and faster vram

Yes, I would like to see another video of this nature for the higher end video cards as I am in the "bubble" for purchasing one. I must admit that this was a rather well put video and I was truly amazed as to the outcome. Well done!!!

wow this is super informational for consumers!! can you do more of this kind of stuff with the other cards of that generation too please?

Really, the only people who should be buying the 3050 6 GB are those who are building on an extreme budged with something like an office pc that doesn't have any supplemental power, in that scenario, it's a great option, for EVERYTHING else, almost anything is better.

Nvidia got greedy...

Fantastic video. Could you please do a follow up with the 80/90/Titan class cards? Thanks in advance 😊

Yes, would love to see more videos about this sort of thing!

would buy a 4060ti - If it had a min 192bit bus and 16gb ram at about 300$ usd problem is the standard 4060ti has 8bg ram 128bit bus at 440$ where its cheapest (my region)- thats a shit card at a shit price. not that anyone should buy a card with 8gb vram in 2024, thats just not enough anymore.

so we just wait for the next generation...

The data was spot on mate . Keep it coming plz , very very interesting to visualize with these graphs what is in most ppl's minds. well done .

Appreciate the work that went into this analysis! Provides a nice perspective for the debates the community has been having for the past year.

Every 10 generations is the sweet spot.

So dont buy cards. Do you think prices will stay the same if lets say Nv sales drop even 40-50% for a month or two? A product cost is what people are willing to pay for it, nothing more and nothing less. If enough(from Nvidia and sellers point of view) people are ready to pay 2000-2500 for 4090 its going to cost 2000-2500.

The video on pricing I have been waiting for years. Great analysis thanks.

Really enjoyed the content. Please continue making similar videos in the future!

As a corporate guy I would read it as "don't make too good products because that raises expectations".

I’m afraid that this is an irreversible trend. Remember when the first $1000 smartphones got released and everyone on the internet said lol nope? Flagship phones are still $1000+ today and sells very well. I really really hope this is reversible for pc parts but I’m skeptical.

That's a clever way to analyze the price increase vs expectations. Thanks, that is enlightening!

Gr8 video ! How about a "power consumption @ video playback" (UKposts, Netflix, VLC etc.) comparison of current GPUs with updated drivers ? * Emphasize on "video playback" as most people confuse it with idle or gaming consumption !

Damn, algorithm knew exactly what I wanted to see. Never been this early before.

The GTX 1070 situation is PEAK nvidia rip of mentality back firing. Everyone who had a 1070 saw how good the 3060ti was, saw that the 4060 ti is the same crap, and went and bought a used 3060ti for much cheaper, thus saving money.

As a consumers we can send the message not buying at those prices. This would be nice to do with the CPU market as well

Amazing video, crazy amount of work, well done!

Lost interest in gaming due to inflated prices of these GPUs might as well spend my money on other things TBH.

F You Nvidia - Linus Torvalds

Please more of such videos! We need to voice our disappoval of those high prices!

Thanks for sharing the insight. Would be great if you can make buying suggestion based on current pricing

but but, the shareholders needs another mansion to brag about - they need more stuff

Everyone moaning that 3xxx series was cheaper compared to 4xxx... How many of you managed to get 3080 for less than 1200 usd/eu in the first two years of the release ?

Phenomenal work!!

Great video, really interesting data you provided 👍

I really liked this data and the graphs were great and easy to follow and understand. So yeah, I'd want more if it isn't to much hassle, but you seemed to like diving down into this, and some interesting data came out of it. Although one would think that these companies would have already have this data, but if not, then you have definitely giving them something to think about in the future. One can only hope that this would make them see how things should have been.

I just think that the amount of research and work that has been done to put all this data to us once for all in a single video... deserves a like, a comment, a share..and more! Well done and thank you.

Superb work hardware unboxed. Love this deep dive

Definitely interested in the upper end of the market. Also would be interesting to see this data with the prices corrected for the ‘onshelf’ prices rather than the RRP (from memory there was a site tracking sale prices of cards) as particularly the 30 series suffered from not existing at rrp for much of the world

Great video! I'd like to see the analysis for the high end cards.

Would love to see this for the high end of the market and the idea of a video that would put into context expectations for future releases value/performance ratio would be a interesting video

this is such a dope video tim, thanks!

This kind of analyses is really interesting. Please do more :)