BAD NEWS! RTX 5080 ONLY At Launch & 5090 Delayed? BIG SPECS UPDATE
HTML-код
- Опубликовано: 15 июл 2024
- Whokeys 25% off code：RGT
Windows 11 Pro $23,2:biitt.ly/581eD
Windows 10 Pro $17,6: biitt.ly/9f0ie
Windows 10 Home $15:biitt.ly/XYm9o
Office 2016 $28,8:biitt.ly/cy7Vb
Office 2019 $48.6:biitt.ly/YInvw
www.whokeys.com/
BAD NEWS! RTX 5080 ONLY At Launch & 5090 Delayed? BIG SPECS UPDATE
There's bad news in the air, as if recent leaks are true Nvidia will launch the RTX 50 series of GPUs with the RTX 5080 ONLY with the RTX 5090 allegedly seeing a DELAY. But, we do have something to brighten your spirits - a BIG specs update for RTX 50. Join as as we examine not only the rumours that the RTX 5080 will launch FIRST, but we also discuss WHY Nvidia may have chosen to delay the RTX 5090 flagship graphics card.
Plus, we also have a RTX 50 hardware specs update, specifically we have new info on the SMs, memory configuration and MORE. With RDNA 4 apparently skipping the high end this gen, it's definitely going to be interesting to see if/how Nvidia changes up their strategy to compete - especially in the highly competitive mid range/budget graphics cards market.
Subscribe - ruclips.net/user/RedGamin...
Join our Discord - / discord
Tip on PayPal - paypal.me/RedGamingTech
AMAZON AFFILIATE LINKS
US - amzn.to/2CMrqG6
Canada - amzn.to/2roF4XO
UK - amzn.to/2Qi6LMS
GreenManGaming Affiliate link -
www.greenmangaming.com/?tap_a=...
SOCIAL MEDIA
www.redgamingtech.com for more gaming news, reviews & tech
/ redgamingtech - if you want to help us out!
/ redgamingtech - Follow us on Facebook!
/ rgtcrimsonrayne - Paul's Twitter
/ redgamingtech - Official Twitter
0:00 Start
0:02 Today's topic
0:50 WhoKeys
1:41 RTX 50 update - 5080 first?
7:47 RTX 50 specs update
Royalty-Free Music - www.audiomicro.com/royalty-free-music 
Игры
Can't be delayed if it's not even announced
When it comes to Nvidia, when it is announced, it will be available shortly after. So, what you said makes no sense. If Nvidia announces a card, then it hits the shelves. There is no time to be delayed.
Every single generation l, all the way back to the Riva, this has been the case.
@@asmongoldsmouth9839 yeah this is correct. Availability at launch is another story though.
OP meant delayed internally, vs leaked rumors of pulling ahead release.. "on schedule" would be NEXT YEAR from what we've officially heard.
Some random announced it on Twitter.
@@aaronriggs4430 No dude. It STILL doesn't work that way. ALL these chipmakers that rely on other companies KNOW things can change and also KNOW the market can change.
So do they have an internal map saying when they plan, that is a soft plan on when they want to put something out? Of COURSE. Do they say "WE HAVE to get this out at this date", absolutely NOT. So that internal map is never fixed in their minds as hard dates. In THEIR MINDS they don't have fixed dates, until they KNOW when their products will sell, and can also be manufactured, and this can also mean not compete and kill the sales of products they need to get rid of, which they KNOW can happen.
To suggest otherwise is saying Nvidia, Intel and AMD don't know WTF they're doing. Yes they do.
It's only the rumor mill that puts out these dates, not the companies. Companies don't leak these dates. When companies know when they're going to launch they announce their products, or they sometimes want to talk about a product line well in advance but don't give firm dates. They'll say "at the end of the year", or "early next year". Or, if they actually know their parts will be in production and are indeed ready to release them the statement will sound more like "near the end of Q3" or things like that.
And this is why I never pay attention to the rumors on release times, because they ALWAYS fail to consider all the factors that go into a company such as a chipmaker putting out a product, and are almost ALWAYS wrong.
So, the problem is the leaks, not the companies who don't have hard targets until they're close to when they can launch and know they have the capacity with the semi-conductor company.
Either way I'll never spend 4 digits on a toy, only if the 5080 is below that it will be on my radar.
inb4 "sToP bEiNg PoOr"
I think the upgrade cycle is getting longer and longer. 5080 might last 8 years
@@delatroy so $999? :D
@@korinogaro if we're lucky.. Nvidia making so much money on AI.. no need to give the peons any value at all
i guess you'll never own a sports car, or even a small boat
if 5080 is 16GB, which is most likely gonna be the case, sad day.
Yep gotta wait for ti, super or ti super. Lol
@@krazyd0nut404 maybe they will use 3GB memory modules instead of 2GB... >.
so? its gonna be GDR7
They will make the 5080 go with 16gb vram. Reason is simply that they need to hold the 4090 in a good place with 24gb. Otherwise they will be destroying the worth of their own product
@@Full_Tilt_Joker i thought they are ramping down production of ada lovelace, well at least for the lower tier cards. but if they keep making server grade cards with lovelace, they would have lots of defective inventory that would otherwise go towards lower tier cards.
i'd assume they would put full production into blackwell by now so they can have plenty of server grade cards to sell to AI companies at a premium.
That means, they will test the market with "5080" (or, 5070 in reality) and leave the 5090 (or, 5080 in reality) for 2025. That, what should be the REAL 5090, will go for more profitable AI chips. Gamers won't ever see the full dye.
It's because Ampere and Lovelace sold terribly. They're still sitting on a decent bit of stock. So launching a high msrp card is to "encourage" people to buy older and cheaper cards.
@@26GuenterThey will "encourage" me to buy AMD instead if that's the case.
@@GoldenSWAMD RX 8800XT for $500 or Nvidia RTX 5080 for $1300? Yes, that would be a hard decision.😂
@tringuyen7519 depends if you want a GPU that is like a boosted 7900XT, or you want a 4090.
Clearly not the same target
If it beats 4090 its not a 5070.
Turns out they can't keep the 5090 from melting their power connectors and they don't want to admit what a major f-up the new connector is. Just my guess.
i like your guess.
The power connector situation is OVER BLOWN ! One electronic repair shop out of how many cards in population??? And that repair shop admitted they have a contract with CABLEMODS who had to recall all there aftermarket 90 degree angle adopters .
No other repair shops have come forward. Granted most people should still be within there warranty window and should RMA there card anyway for a replacement not paying a 3rd party repair shop .
Way over blown issue
@@bigdaddyt150 They have a 10% safety margin on that connector from first approximation. Any electrical engineer worth his degree will tell you that's cutting it close at best. The OG connectors are good over 150% their rating. They're possibly overkill.
also you don't run that close with a connector without having tighter tolerances on the connector and redundant latch connectors.
@kaseyboles30 I get that !! But you do know they have made millions of RTX 4090's and we all are basing the failure rate off one Repair shop that's on the west coast who is also contracted with cablemods who admitted to making faulty cables and adopters for that particular power connector
@@bigdaddyt150yet when y spend over 3000 or 2600 y don't need a cance to live in the cold since home burned down by the connector 😂
I wouldn't mind the 5080 being significantly slower than the 5090 if it's significantly cheaper as well. But if it's still around 1000 usd and with 16 gb of vram then it's a hard pass.
256bits suggests 16gb, the 5090 with 512bits probably has 32gb
I guess they want the 5080 to be below the 4090 in some specs so they don't have to sell the 4090 for half the current cost
I cannot imagine nGreedia giving us 32GB on the mid range PC desktop, and yes, despite being expensive the ‘80 cards are now upper midrange. Not for another 2 or 3 years, if they don’t pull out of the gaming market by then.
@@tonep3168Yeah Nvidia did kill the 80ti this generation. Both companies are corrupt but Nvidia takes it a step further by tarnishing their own products potential.
Oh man I was hoping for 20gb or 24gb in the 5080. Is the 256 bus width for sure?
It could still be 24GB if they use the higher capacity GDDR7 modules that were talked about.
Personally, I think it's just another marketing scheme and shitty business tactics to launch a 80 class card and count on ppl just being thirsty for another card so we'll sell them the weaker one first
If the 5080 is 20gb VRAM and priced right I expect it to sell more then the 5090 would since there's no way in hell the 5090 will be under 2,000 USD
If the 5080 is $950 and matches or beats the 4090 that would be very tempting for those of us on much older cards.
And with 20 GB of vram, yes
dreaming if you think ngreedia would ever sell that cheap
@@matiucosmin3112 it's 16GB
5080 will probably be $1000 and be 5 to 10% faster than the 4090 in 1440p in non Path Tracing/heavy RT workloads, but in 4K with heavy RT, the 4090 will probably still be better by around 5%.
just a guess
They won't destroy their own product. If the 5080 comes for under 1000$/€ with more than 16gb vram you can flush 4090s in the toillet.
That won't happen
Looks like the 5080 is gonna be aimed at the just under the performance level that avoids the China export ban. So just another 4090 D or perhaps even a bit weaker. While the 5090 will be a paper card until the professional market demand is saturated with the GB202 chips.
This makes sense
hey, ill take a 4090 for $999, as long as its cheaper than the 4090 its a win
If the 5080 has 20gb of VRAM and is priced like a 4080, people will buy it. The 16gb on a 4080 now is just planned obsolescence while not everyone wants to pay 2,000 for a 4090
@@jcdenton7914 4090 is a waste of a card in my opinion.
Because it's still too expensive people who own 4090 should sell this card asap before Nvidia launch the 50 series and when they do it's too late to sell it. Because 50 series especially 5080/5090 will be raytraycing monster GPU's and will also able to handle pathtraycing which is enhanced version of raytraycing. Something even the 4090 is struggling with everything maxed out only 35-48 fps tops, 5080 will push this to 60-85 fps and 5090 above 100+ fps and with each update it will only improve.
We've had the introduction of the "RTX" branding, then we got the mining boom, then the global component crisis and now the AI hype. All of these things combined with some inflation has caused a massive price hike for a GPU. The gaming community deserves a break next gen with affordable, high performance gpus.
They saving those chips to pump out ai cards. Just greedy that’s why. We get the scraps
There in business to make maximum profit , how is making a business decision being greedy .
Diy PC is a tiny tiny market full of entitled whiny children whereas nvidia is a top 5 sp500 company which is deemed of national security important by the US govt. Yes, they are in the big leagues now and don't care about the toddler wiffleball league
Every company does this though. In the TV market all the displays are scrap to Micro LED.
Let's be honest, you only deserve the scraps.
This is actually very smart on Nvidia to do this. People will double dip when the 5090 comes out
I will just stick with my 4090 and get a 5090 when it comes.
No they won't. Why would they go from 24gb to 16gb vram unless they're morons lol. I don't see the 5080 being much more powerful, if at all then the 4099 besides in ray tracing. This is more to make up for the terrible 4080. The card that was priced way out of its price range. They're not going to put out an 80 series class for over $999, so putting out a cheaper 80 makes them look better for giving us a cheaper card with 4099 performance. Even better if they go with $899. The specs suggest it's going to be less of an upgrade then we usually get but in sure they've done a lot to make ray tracing a lot more performance so that's what they'll market it on
This GB203 looks like a 5070 compared to GB202. I think I'm gonna wait until late 2025 before upgrading
Anything regarding power delivery (connections ) ?
If it's not a 90 card - I'm not interested...
they can sell ya GB203 and name it to 5090 like 4000 series "4080"
What most people are interested in is a card that beats/matches the current 90 card, for half the price and amd might just bring it to the table, this is why nvidia is making all the noise.
90 cards are waste. unless its used for A.I..
@@a2raya772 such as me. Otherwise i pretty much agree.
The last PC build i recommended was with a 7900 GRE.
@@25myma yet stats shoe most people buy 4060 :D
yea it is true, also in other products. people look flagship and turn off brain to pick something for their budget.
Generational increases have been 60% increase in cores from 3090 to 4090 with 30% more native 4K fps, similar to that of the 3080 when added to the comparison, 20-25, 35, 45-50 respectively. We can expect 70 fps native 4K, perhaps even 80-90 fps with nvidia focusing on Ray-Tracing and these examples of framerates being 4K max settings cyberpunk with psycho RT.
I've said this for over a year now and will continue to stick with it.
No chance that I will replace my 4090 with a 5080, knowing that a 5090 is on the way at some point. That said, if nVidia delay that gives AMD a chance to show us if they have anything that can compete in the same class, although I doubt it.
Since you have a 4090 skip a upgrade this upcoming gen.
why tf would you even buy any 5000 series gpu?
Won't even upgrade the 3090 to the VRAM lacking "trailer park sku"
@@lorsch.Same. I have a EVGA 3090 and I won't upgrade either if the vram is lacking
@@suparibhaubc he clearly wants high end performance which amd literally cannot do to save their life 😂 their bread and butter is console hardware and cards for poors
This release schedule rumor coupled with the new specs suggests to me that GB202 is just two GB203 dies, like nVidia's doing with GB100. The added manufacturing complexities could push production back a few weeks, letting the 5080 launch sooner.
If only the rtx 5080 has atleast 20+ vram, ill pull the trigger to buy one, pair with ryzen zen 5 and 4k monitor.
I'm not sure how a product is delayed when it even hasn't been announced. Nvidia hasn't publicly said ANYTHING about these products.
Most companies set internal target launch dates months and even a year in advance for their products. Just because a launch date was never announced publicly doesn't mean it doesn't have one.
It's delayed based on an imaginary date in people's heads.
@@03chrisv See the other comment.
Having a road map, which are almost NEVER exact, and are mostly created because these are public companies that have to talk about the current state and what they're planning to do doesn't mean ANYTHING is set in stone.
So, TSMC could have a delay, Nvidia can't predict that but if they need Apple and a few other companies to be off a node before some production line can be open for their die, it's going to delay when an Nvidia line of products can come out or it's a paper launch and these companies will NEVER talk about TSMC publicly other than to say the polite things companies say so you don't piss off the company that allows you to make your products.
They could evaluate the market and say they should wait a couple months for many reasons.
So no, there is no hard dates that these companies set. There is a road map published to show their board of directors and that's about it. After that, they have a general idea about when they want to launch.
You know how many times Paul has had to change what he says about dates???
Do you think these companies don't understand what it takes to launch a product???
You don't think these companies can read a market????
Go back 2 years ago and AMD and Nvidia BOTH predicted what their sales would look like over the two years that just passed. They predicted when their desktop sales would start to pick up and that's about when it picked up, so CLEARLY they have an understanding about dates and what they should do and no product is ever delayed unless it's been TALKED ABOUT by the company.
I'm not saying they don't make internal plans. I'm saying they know their business so those plans are NEVER set in stone and that's why they don't give targets to the public until they KNOW they're going to launch at a particular time. Zen 3 was talked about well in advance. It came out when AMD said it would. Zen 4 wasn't talked about well in advance. It came out when AMD said it would.
@johndoh5182 That huge rant for nothing. You literally just reiterated and expanded upon my comment lol. I never said anything was set in stone. A target launch date is just that, a target to aim for. Nothing more. If a company wanted to aim for a September launch of their product, but unforseen factors pushed it to November, it's still a delay from the perspective of the company regardless if the public ever knew about their September launch target or not. Considering Paul is reporting on INSIDER information it makes sense that it would be reported as a delay because from the perspective of the insider it is in fact a delay.
They orchestrate these leaks and rumors to manipulate the public into buying their older stock. They have teams of former CIA psyops project heads designing these whole campaigns. They play us like a fiddle.
If it’s not chiplet and we know chiplet is the future, would this be considered an intermediate product if they plan on a whole reinvention with chiplet design next. Is that why it’s being rushed to sale?
Dam remember that in 2020 we had 4 cards, 3090, 3080, 3070 and 3060ti. Pray in god for the 4080 being less than $1k
So the 4080 / 7700x was my first GPU / CPU combo in 2022 and I would love to upgrade for the 4k 240 OLED monitors, but I can't justify it during the 50 series and Zen 5.
Is the 60 series and Zen 6 a better step and time frame to wait in between PC parts? 2 gens ...
Wait for Zen7(DDR6) and the 6080.
Your 4080 and 7700X will hold you well for quite a long time, specially bc now every game has DLSS and the implementation is only getting better, keep your 120FPS monitor as well.
@@MaxIronsThird 🤔 that would be AM6 most likely, no? And an entire platform change.
@@chrispreczewski yes
@@chrispreczewski Every new platform like AM5 or AM6 is a bad start. Because the first gen of CPU's for that platform are not fine tweaked and will cause problems that's why you wait for the 2nd generation of AM5 and if you want AM6 you go for the 2nd gen AM6 processors which is Zen 7.
So take my advice and avoid these issues.
Personally I'd go with 5080/5090 with 9800X3D/9950X3D.
Or buy 7800X3D or latest 9700X and sell your 4080 and get 5080.
Skip Zen 6/RTX 6000 and go for Zen 7/RTX7000 🤫
It seems relatively obvious to me. The reason that NVIDIA might launch the 5080 first is because that's the highest likelihood that it's going to get the earlier adopters that just want that new feature set and performance and aren't willing to wait. Those that want something that is simply good enough. If the 5090 is significantly more powerful. They would likely go that instead.
I'm super excited to get that 5080! Thanks for your cutting edge info. And... damn, my guy! Haven't seen your vids in awhile. You got swole (no homo)! I guess you changed your diet and are working out a lot now!
To me it is still a dual GB203 fused into GB202. This explain the doubling of specs differences. Also what are the chance it could be Titan instead of 5090?
And finally the memory bandwidth is corrected to 512 bits bus.
All depends how Jensen feels 5 minutes before the show 😅
Man's getting buff!!!
So $1300+ and probably 16GB vram? Table scraps for sure. A bigger curve ball would be Nvidia announcing that both regular 5090 and 5080 use GB203 and use GB202 for an AI focused 5090Ti.
Nah the 90 class was always a Titan class and they never dropped those GPU’s a tier by the chip. That is just a dumb take to make.
Is 5090 and 5080 even going to run properly on an x670 pcie gen 4 gpu slot?
I'm curious (hopeful) about the efficiency of the 50 series. Hopeful for less power use as that 4090 is a beast.
Remember man it’s performance per watt. So if a 5090 was %55 faster than a 4090 but pushed the same amount of power that is efficiency.
There is a big gap there - definitely space for another GPU in the stack.
I wonder if they’ll do something like:
GB202 RTX 5090 $2000 (512 bit)
GB202 RTX 5080 Ti $1500 (384 bit)
GB203 RTX 5080 $1000 (256 bit)
the 5090 will have its bus width cut down to 384bit(24GB/no 3GB modules yet), there won't be a 5080Ti from the GB202 die, probably just an OC'ed and less cut down GB203 with 3GB and faster modules
I'm just waiting for a *RTX 5050* 70W TDP
Laptop?
it will be launched by 2026
As a 4090 owner Flight Sim VR runs fine and so does Cyberpunk. Most games come from consoles and run fine at 4k/60 - 120fps. 5090 is nice but I don’t feel the thirst for the power. Although i may end up upgrading just because it’ll be the best time to sell the 4090 to put towards a new card.
Ever since foveated rendering became a thing my (OC'd) 3080 10Gb also does a good job in DCS VR - reading the power draw figures VRAM is not a huge issue.
So even I do not feel a hurry to upgrade, if the 5080 will be of poor value, I can wait 1-2 more years. With a 4090 I'd definitely skip.
This a awesome update thanks
The 5090 better have 2 12VHP connections. If 450 watts is melting connectors, 600 watts will cause fires. But if they use two…
i'll only upgrade if they add more VRAM to 5080.
hahaha you tell em!
Love the Alien T-shirt, just finished re-watching all of the movies. First three definitely the best.
This sounds more like a 20-25% upgrade of 5080 vs. 4080.
If it is still priced at 999$ (and also available at that price) this sounds like a decent upgrade and the card that still makes sense for the broader market.
If street price goes back to 1200$ then it is rather meh.
5090 sounds like a huge monster above 2K and alsoi very power hungry. While performabce is great I dont think I want to spend so much money even if performance is great and power consumption will be too high for every day use even when undervolted.
old news, gives more time to save 5k
Releasing after they have them would be nice.
They have these things called a hairbrush ....
Value is further down the product stack - So over the year's working up from 50 to 60 then 70 to an 80TI class card, I'm dropping class to still faster but cheaper & using less power - so depending on what the 5000 series offers, for me as long as it's a bit better than my current 3080TI, cheaper & lower power, I could be going for a 5060TI class - will have to see - but I'm done with the high end, it just costs too much for not that much more 👍👍
The 5090 better come with DP 2.1 and Next gen 2025 monitors better have the DP 2.1 port
Seriously. The fact that they all released 4K OLED 240Hz monitors without DP2.1 ports is disgusting. They only did it so the whales will buy one now and then buy a whole new one in a year or so when the 5090 with DP 2.1 comes out. The folks that bought a monitor that they couldn't use to full potential because of cable bandwidth, will buy another replacement for it that can.
@@joeshmoe9233 If I'm right 5090 may support DisplayPort 2.1a a newer version with high bandwidth support for higher resolution up to 8K 240Hz or 8K 120Hz so don't worry and if they forget about this AMD and everyone will start making memes with this XD
The 5080 first is fine with me, as long as it has a not-too-insane MSRP and reasonable availability. Thumbs up!
Hard to imagine they'll give better value than last gen when they're utterly creaming it right now on the business side :/
AMD also won't have high-end SKUs so no need for Nvidia to be price competitive at all
A like just for the alien shirt alone!
Honestly would be such a huge bummer if the 5080 comes first, only interested In the 5090 and really hoping it’s 512 bit bus with 32g vram
yea cause you need 32 gb of vram hahaha. even the current 24gb is overkill
@@iequalsnoob overkill? Your just getting into PCs or on a Xbox still?
@@iequalsnoobIt’s not overkill because people do more than game on a PC.
Great, I was waiting for 5080.
5080 needs at least 20GB
Something wrong here, because gap between GB202 and GB 203 is too big...
1. Either GB202 is not 512 bit, but 384 bit die
2. or 384 bit die will exist along with 512 bit die (GB202)
3. or 512bit die (GB202) is MCM and basically 2x GB203
If 5090 is going to jump in performance even marginally, the question is.. will it have the new display port 2.1?
Name the video "Double talk. Extending one sentence to 15 minutes."
4090 power still not settle , how the 5090 can ?
Hope they price the 5090 at 4k or above. And eventually the 5060 at 250$. I would have no problem with this at all.
I predict when RTX5060 launches it will cost what the GTX1080 cost so if you are looking at updating from a GTX1080 for the same price or if you increase your budget by inflation you will be limited to RTX5060 or RTX5060ti. Nvidia most likely delaying RztX5090 as get more from AI and no AMD or intel competion and not China.
How many gpus can you Bench 💪 not benchmark.
120 GPU'S (4090's)
If there is no high end competitor 5080 will be the flagship for the 50 series cards
I've been saying this for weeks since we got the rumoured 96sm specs for the 5080....but nvidia knows selling an 80 series class above $1000 is a no go. So instead of slimming their margins they're just going to give us a less powerful upgrade then we would normally get to save on costs. So my guess is a 5080 with 16gb, raster on par with the 4090 but maybe 20-30% increase in ray tracing since i think they'll really Want to push for ray tracing to keep ahead of AMD. and the price will be $899-$999. For $899 with 4090 performance I think it would sell very well even with "only" 16gb vram and less of a jump then the usual 50% increase we get and they'll advertise how much better it is at ray/path tracing. At $999 it's a little less appealing, especially for 4080 owners but I think the bigger ray tracing uplift would convince many. That's nlmy theory anyway. I'd certainly upgrade from my 3080 for $899
So rtx5080 same speed like rtx4090 or 10% weaker price assumption is 1200$
well the only way this is good is, if the 5080 and 5070Ti launch first, and only if those cards aren't even close to 900 dollars and even if they are they both better be faster than the 4090 or as fast in gaming. So basically if the 5080 is 699 or 799 like the 3080 launch and the 5070Ti at 599 or 499. If they did it this way, it actually wouldn't be that bad as the midrange crowd could get a killer card this year and the high end crowd wont have a to spend as much as they did last year, and AMD might actually have to compete even further.
However considering Nvidia Being Ngreedia these days, I don't even know if that possible, but we didnt even think nvidia was going to almost be-rid of the Ngreedia RTX tax when 30 series came out so i guess you really can be a bit hopeful especially when the 4080 just wasnt sold as much as the 3080 considering the launch price as well.
You’re kidding yourself to think that the 5080 will be $800. Try $1300! AMD RX 8800XT will probably be only $600 or less.
@@tringuyen7519 yeah that's what people said before the 3080 came out because of the aledged "RTX tax" at the time from 20 series, it turned out to be 700 dollars.
I'm not saying it will be, I'm saying its the only way this could be good, if its going to be 1000 dollars at launch and its the only card to showcase and i know Nvidia could do it if they did it before, if they don't not many will want to buy it, only the few who buy the 4090 and 7900xtx at launch will probably buy that thing at 1000 dollars, or if you're an Nvidia fan but who knows at that point.
Lol a 5080 for 699$
So what are you dreaming??
Price drop for 4090's to 300$??
@@Full_Tilt_Joker yeah and the 1080ti was 700 dollars. The only reason were at this point with graphics cards is because people actually paid 1000 dollars for a freaking 80 class card with 20 series.
30 series almost turned that around with the 3080, but the only ppl who bought it in bulk were miners and scalpers who bought them up at 6-799 and sold them for double and or triple. Ngreedia is a trillion dollar company, they could price the 80 class at 799 if they wanted to, but they probably won't, which means they've limited divided the high end even further from the midrange, and with the 40 series the 70ti was reaching that utter limit it should've never been at.
Also the 3090 was basically a gaming titan card, and was priced as such, that's why the 3090 was pretty much 300-500 dollars more than the 3080 at launch, for a card that could pretty much do it all, you got what you paid for, it made sense. The 3080 was the stop gap, it could do what it needed to do, but it wasn't overkill and it could do 4k over 60 fps and not break a sweat, but for much-much less at the time... But because of crypto miners and scalpers who just bought them all up, and with everyone wanting a new computer, the market couldn't meet the demand and many were left out.
The 3080 was not 1000 dollars msrp and not even 900, far below that. Does no one even remember Nvidia priced it this way? Like if they could do that coming off the most scummiest GPU launch ever that was 20 series, I think it's fair to say Nvidia could pretty much do anything at this point, people thought that prices for the 30 series were going to be worse, but no, it was actually more bang for buck than the 20 series ever was.
great britain 202 :D
Wait 2 gens from 4090 or go amd sikilar speeds my 7900xtx to 4090 for a lot less money
I got a 4080 currently. I think this makes sense to wait atleast 2 gens before upgrading to another gpu.
My thought on the 5090 is they're possibly creating (yet again) their own 12v connector instead of using the new one, or its going to use two of the new 12v connectors to circumvent the obvious limitations of a single connector. There's a lot of uneducated people out there who think the 12v connector issue is overblown, and they'll be proven wrong once we see the 5090 and how it's powered.
I'm waiting to see what AMD announces . I'm really curious despite the rumors see if they will drop a high-end card
Hhmmm, they did say a few months ago that they won't be competing at high end anymore.
@@maydaygoingdown5602 AMD never said that. Rumors
@@BigAndTattooed No I think they actually did. I remember Adored TV doing a video about it a good few months ago.
their top card will be a gimped 7900 gre. thats already been leaked
@@iequalsnoobleaked it must me true then you got me there.
I expect RTX 5080 is 30% faster at 4K than 4080 (raster) and 50% faster in RT - price at 1200$. Looks OK for 320W GPU.
If so it will be the first time a next generation 80 class is slower than a last generation 90/Titan card.
If it doesn't beat the 4090 by at least 10-20% then it's a huge disappointment.
@@ZackSNetwork it's possible. Because 4090 was the biggest leap in history and there's a huge difference in SM numbers between GB202 and GB203. None of 80 class cards achieved 70% uplift in a generation. I think this time we'll get closer to 4090 cheaper and energy efficient gpu. It's not that bad.
@@jaxonswain3408 it will be 25procent better for 899 bucks they will throw them away and laugh at the 4090 buyers who can sell there card for 500 bucks now rofl
You can see in the B200 design that GB20x will be 1 BW die at 8x0 mm2 so said 132 SMs. Within that die there are 4 compute complexes (vector or parallel arrays u chose) you can see them. If GB 203 sticking with 400 mm2, similar 4080, will have 66 SMs. The 200 mm2 die will then be GB 20x, cut in half again having 33 SMs. GB 202 the full 133 SMs? Well, maybe 5090ti or BW Titan. Albeit using only 600 mm2 of an 812ish mm^2 single die would be a lot of wasted silicon area for GB 202. I'd say GB 202 needs to be laid out again for optimizing its die size unless Nvidia swinging from TU desktop design generation to Ada mobile design generation and back to BW desktop design generation offers up the whole BW die at 812 mm2, I can't exactly see that for GB202, maybe GB 200ti, but at 4NP depreciated process the cost to fabricate 812 mm2 of BW should be no more than 602 mm^2 area for Ada on the leading edge of TSMC 4 'cost wise' with 4NP NOW on the trailing edge 'cost wise' which results in a lower dGPU cost to Nvidia and price to AIBs offsetting the extra cost of board layers and traces required for wide bus and more VRAM design as a desktop and professional workstation card. mb
192 SMs for BW 202, sorry, 5090 will not be x2 BW dies at 1600 mm2 of silicon. There will be no stitched together consumer gaming products. Speculating GB 202 at x2 BW dies (192 SM) subset of B200? x2 133. Sorry, not at a price any gamer would be willing to pay and compiling for that behemoth mb
I will probably buy a 5080 to replace my 1080 but no way am I going to be an early adopter. Wait for the competition to heat up and
Plot Twist: GB202 is only for AI cards and GB203 is the new 5090.
I will answer on a perspective for who doesn`t has the budged to get a 4090 or even a 5090, the RTX 5080 if is the same performance of 4090 and if is not much pricier for me it`s enough, because I was unable to buy 4090 and I will be happy to get an 5080 instead but sad to not get 5090. I am Brazilian, things here is way more expensive because aside the price of the original seller we need pay taxes to import the card, and we are talking about the same performance, but this card could be faster than a 4090 and the 5080 I will able to install and forget instead of worrying if it will melt.
I want the damn 90!
probably churning out 5090 gpus with a lot of minor failures, so they become 5080 while they build up supply of good 5090 chips.
My guess is Nvidia is putting out the 5080 because AMD has said it's not competing at the high end. If Nvidia can still beat AMD at the 80 level, it's a more cost-effective win. And then, in the off-chance that AMD was sandbagging and brings out a high-end card, Nvidia can adjust the 5090 accordingly.
I’m not a super leaker, but Mark my words Nvidia announce the 5090 to 5080 and the 5070 and they’ll stagger the launches just like they always do
how much will rtx5080 cost ???
I suppose it will cost over 1000$ at least .
1000+ and a bit of your dignity
Way more than $1000
A40 = Ada L40 = Lovelace I bet it will be B40 = Blackwell
Hopefully they won't price it more than $1000. It will be another disaster if they increase the price of 80 class in this generation as well.
$1,000 is still way too much, and people buying it at that price is why it's become normalized. The 80 class card should be no more than $700 at most
My expectation for RTX 5080 is it will match or near the level of the RTX 4090 at around 350W to 400W range, possibly set at the China AI ban limits. Nvidia will want to avoid 450W or higher power consumption.
As for RTX 5090 with the details of a larger monolithic die, this means more power. When it comes to the power consumption, Nvidia wants to keep that concern buried. Nvidia will want all the problems of the RTX 5090 being the US-China AI ban or production concerns of supply and demand between professional cards and consumer cards. Nvidia would not want the problem being having to fit 2 12+4 pin connectors onto the video card to get around the melting connector issues, which will require PSUs that can support 2 600W connections or 2 PSUs. Anything unveiling the power demands of the RTX 5090 will essentially be bad for Nvidia.
They will release new low VRAM overprice GPUs.
Honestly I think RDNA 5 might beat RTX 50 series in raytraycing.
But not the 5090 however next gen XTX may be on par w 5080.
Keep in mind RDNA 5 is powerful but not ready pathtraycing.
Even if it's able to do it will struggle same as when RDNA 4 came out.
When RDNA 4 came out the raytraycing performance was not great.
Same thing is gonna happens again, instead of raster now RT is great,
RDNA 5 is not ready for PT at least not as great as Nvidia's 50 series.
5090 might come in 2 variant a 512 bit and 384 bit it could be Ti card.
5090 512 bit 32-48 GB $2199-2499
5090 384 bit 24-28 GB $1599-1799
5080 256 bit 16-20 GB $ 999-1199
Note: the 512 bit 5090 may be a limited stock/limited edition.
I have a feeling that with no competition from AMD, Nvidia will launch the 5090 at 2000, reduce the 4090 to 1500, with the 5080 in-between the 4080 and 4090 in terms of price and performance. Then lower the price of the rest of the 40 series to compete with RX8000, and Nvidia wont launch the rest of the stack and the refresh of the top end, until AMD is ready to compete again with RX9000.
Yes terrible news indeed, anyway...
You can buy the 5080 for $1300. Or you can buy the PS5 Pro with 8 core Zen2 clocked higher, 60CU RDNA 3.5 with better RT, & 300 TOPS NPU for PSSR for $600!
@@tringuyen7519 bro get your 1080p blurry as* consol brain dead as* out here. your not gaming.
I dont think people really give a shit about the 5090 launching. Like any of us are gonna be buying the overpriced 90 variant crap.
some people just interested to see where 5090 will be even if they only buy $200 GPU.
It's about what we lost over time for some.
Me personnaly, i like what AI.
The more VRAM the bigger and better model Can run and the faster it train. Huge time saver. I value my Time much more than my FPS (and my money at this degree)
@@arenzricodexd4409 Yes of course, even years ago when i was a poor piss peasant who couldn't even buy a 50 variant GPU always loved to see it. But i also dont care about it because ill never buy it. I could never justify spending 2k on a damn GPU. My entire PC setup costs as much as a 4090 and i am playing games in 4k just fine. Such a waste of money.
@@lefourbe5596 IF you are buying it for work related tasks and not solely for gaming, then it makes sense. If its purely for gaming then you are an idiot. Or very wealthy.
@@cleny217 waste of money if you only game on it. But those that use it for productivity and making money from it the 4090 are worth it since they did not need to fork a few thousands more for cards like A5000 or A6000
dont give nvidia your money. wait for 90's launch
I could consider the 5080 if it is less than a grand, is at least 256-bit 16GB and uses three 8-pins instead of that stupid horrible 16-pin connector.
No chance
@@ladis79 I'd like to think they recognize that people just don't want to buy 80 class cards at more than a grand by now and would design them to be sellable for that price going forward.
Tho using the proper power connectors instead of the ugly melty boi - I have a feeling that will take the FTC telling Nvidia that they can't force AIBs to use the trash because it's unfair competition to prevent them from offering more desirable product.
@@cracklingice unfortunately there is no real competition from AMD for upocming high end and I am affraid that greedy NVIDIA do not have any reason to keep prices low...But it is all pure speculation, let´s see and maybe you are right🙂
@@ladis79 80 class stopped being high end when Nvidia released 90 class. 80 class cards are just the 70 class with a name change - particularly in the 40 series where the one they call 70 class isn't even properly 256-bit and 16GB!
"Jensen" Introducing the RTX 4090 connector melter dig out the marshmallows edition, Warning beware of toxic plastic particles may be on the marshmallows do not eat for your own safety, Nvidia™
Beard suits you😄
How is this bad news when its going to be the most expensive card
Its already bad enough that youtubers act like the 90 series is the most important card when its gonna end in the towers of the few
Looks like nGreedia is up to shenanigans. I just can’t see the 5090 being the same as the full AI card. Yet, its performance is not going to be much more than 60% faster than the 4090, assuming the 512bit bus is correct. Plus how can nGreedia be allowed to sell a product to China with even 1% more compute performance than the 4090D? So that takes me to the 5080, that would surely not be able to outperform the 4090d, as this card is most likely what nGreedia will want to sell to those rich Chinese that they love so much… It makes me wonder if the 5080 is going to use AB104, as I just can’t see how this product stack is going to work with the Chinese limitations, and availability of the AB102, which is HBM, so that’s why I think the 5090 will AB103, a smaller more easy to manufacture die, and the 5080 and possibly 5070 will use AB104. No matter what, I think us gamers are going to get shafted again, just harder.
just go buy amd crybaby
@@iequalsnoob Come on girl, be kind.
60% would be a big increase and no channel would say that was bad, as long as it was not much more expensive.
like the 1080
5080= $1799
excellent i will start the process for a new mortgage and still wait for AMD
Nvidia trying to sell out their overstock of 4090s.
There is none.
RTX5080 is all I need
5090 delayed...bad news? LMAO! Sure, the vast majority of us are really heartbroken that we can't buy a $2K gpu.
Does Nvidia hate money? - the halo card always drops first.
Makes sense to inflate 80 series prices as the flagship, then drop TI or 90 series stuff later. Crazy mf's will upgrade from 5080 to 5090, buying both. From a capitalist standpoint, they could drop 70 series first if it were faster than 4090 and inventory would dry up.
If the 3080 was a real 80, then 4080 is a 4070 and the 5080 is close to a 5060. This happens when there is little competition, esp above the XTX level. Jensen is using the fact that will not change w RDNA4 to the squeze gamers even more. The 5090 dies will go to 2-3x priced pro GPUs, in worst case until AMD launches a RDNA5. The upper end gamers is getting royally screwed another year or two, since the 700USAD 3080 was a thing. The lower end may prove the same, if AMD does not launch the RDNA4 before next year. As it will take profits from RDNA3 cards, they have little urgency to launch it earlier.
Was the 1080 a real 80?
Let's wait it out, I'd say nonoby would agree that almost any card of this generation was a big throw for either of the two AMD or NVidia.
If the NVidia prices are still outragous AMD has another 2 years, okay maybe rather 1-1/2yrs till NV redoes their lineup again, to show how much they can do with their FSR.
@@Quast Agree, but that is because normal people could buy the 3080 15% slower than the 3090 at $700. Today the 4090 is 25-30% faster than the 4080, but at 1600-2000 nobody has it, and at 1200$ few woul buy the 4080. The 4080s was 3-5% faster but cheaper, but was a plaster on a broken leg.
Looking at the number of cores on the 50xxm 202 and 203 it is a repeating pattern.
@@damianich4824 I did not game at the time.
Only 256bit bus for the 5080 means I'm not buying this gen unless the 5070 also the same bus width. I found that too low bus width can bottleneck in some stuff I care about. I'm not spending 1500$ on a GPU with still such a narrow bus and 16 vram, go $k yourself Jensen.
hahaha go buy amd you bum
5080 16GB, 5090 32GB?
I think so.
They need to keep the 4090 attractive. By keeping the 5080 at 256 16gb they do just that
It will come back the vram. If the 5080 is 10 or 12 vram there is no way I buy it.