• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

NVIDIA Grabs Market Share, AMD Loses Ground, and Intel Disappears in Latest dGPU Update

I clicked on show ignored content to see what you guys were replying to.

I regret that decision. Wow.
 
At least Intel clearly distinguished their rereleased and new products with the Ultra moniker and seem to be avoiding using the exact same numbers, very weird marketing, but relatively valid IMO. Better than the "13th and 14th Gen" Alder Lake parts, clearly rebadged 12th Gen CPUs. Nvidia's puzzles me a bit too, for example, there's the RTX 2050 which is clearly an Ampere GA107 model, basically the same RTX 3050M but with a very narrow 64 bit bus... and they also sell that exact thing as a GeForce MX 250, so the buyer can at least expect to have generally the same ballpark of performance by buying any of these models.

The last big, big oof NV did with branding imo was the GT 1030 with DDR4 and GDDR5, DDR4 just wasn't good enough for that little thing and the performance difference between both was massive. The 3050 6 GB and 3050 8 GB can also be a tad misleading, but in general the 6 GB models don't have a power connector and are clearly labeled to have 6 GB on the box, so that can often be forgiven depending on the context as they make no effort to conceal that it is an entry level offering to the RTX 3050 series. They are all parts for mass market with insane volume, anyway.
I was talking about laptops specifically, nvidias lineup is pretty clear. Yes you can say that they don't correspond to their desktop versions but within the laptop ecosystem it is pretty obvious what you are getting. On the other hand with AMD, a 9910u (hypothetical cpu using amds naming scheme) sounds like it should beating the living daylights out of a 7550u, but instead the 7550u would be 3 times as fast, go figure

I forgot that happened in this thread LMAO

I'm sure my 3060 can hit 160 with DRS on too :laugh:
My 4090 can't, even with DRS, with the game maxed out.
 
....Bought a B580 off facebook market place just to play around with. Actually kinda like it would love to see Arc further developed.......
 
Well, AMD and Intel have only themselves to blame.
Bovine feces.

NV is the dirties player in the market, covert EPP is in full swing and the scandal around that Jesus guy is just a small part that happened to slip out.

Where is 9000 series Asus ROG?

Could the market be about to pop ?
Nah. Last gen stuff by both vendors was available to grab to near MSRP for months.

It's just DIY is not the major player here, ton of stuff is moved by OEMs in prebuilts.
 
Well, AMD and Intel have only themselves to blame.

Intel should have secured more production. And fixed its driver overhead issue. If the driver overhead thing wasn't a problem and gpus were readily available at msrp prices, then it would have done better.

AMD screwed up badly. Their GPUs are prices way too high. I agree with gamers nexus who said it should be 25% cheaper because AMD is behind Nvidia on the features front. AMD playing catchup. I can get a 5070 ti for not much more than a 9070xt. A 5060ti 16gb is also roughly close to same price as 9060xt. Supposedly there are cheap 9060xts and 9070xts but never in stock.
It's obvious, that Intel was riddled y foundry and CPU µarch issues. That's why this undermined. The shadow of these particular issues, wou;dn't let it possible, to make their GPU drivers fine and reliable.
However, these production and R&D issues, might be the very reason, why Intel continued to invest into their dGPU business. These are only speculations, though, but it seems Intel wanted to depart form their past problems, their bane, by making at least one of their branch successful. That's why they went all-in with dGPU business, even despite all the complications. But nVidia and AMD have both helped Intel with their extremely late, extremely slow catch up.

AMD on the other hand, had their heads in their *rses for too long. Their catch-up is never in time. Their lack of potential lost of Consumer/Gaming Radeon division, won't magically auto-regained by Enterprise/Server/AI GPU branches, overnight, if ever.
Also, AMD doesn't have the sreaming services, datacenters, to substitute the lack of powerful consumer GPUs with more or less sufficient proprietary client HW/SW. If they will loose the consumer GPU market, they will never recover.
AMD is still loosing AI market to nVidia, badly. And will do so, even if some very potent ASIC will emerge overnight, and nVidia's superiority will tank heavily. Even worse, because, if nVidia will loose, so will AMD, due to their eternal catch up/copycat strategy,
Their products might be great, and superrior to their rivals (at least at CPU market), but their availability and stock hurt both AMD themselves, and their buyers/potential buyers.
Not to mention their dumb delays, that work agaist, the longer the products has not been on the market. As much as AMD have promised, about the RDNA4 being the "affordable" and accessible GPU generetion. Lies.

Also, AMD should have had the stronger grip, on their own price making, their AIBs, and the retailers, destroying the sales, by making the AMD's and nVidia's SKUs basically idential in price.
AMD should have launched first the latest gen, not wait for nGreedia first and adapt the prices accordingly.
Sometimes I feel that the Marketing and Management idiots from AMD are too incompetent to lift that company .
They sabotaging themselves so much. Even when their products are very good and competitive (like RDNA2 and 4), the Marketing and Management, happen to screw all the hard work, and do it blistering fast.
Sometimes I have feeling, AMD's Marketing and Management are owned by "rivals".
Obviously it doesn't matter if the latest Nvidia product is having problems, on driver or even hardware level, or if it is ridiculously expensive, or if it comes with limited VRAM, or if the marketing is creating a false image of that product. It doesn't matter what we will post here, or what Youtube tech channels, or tech sites will say/write. Nvidia's brand is too strong today.
Was it surprise to anyone? This is the same as with "premium" and dominant company, e.g., like Apple, Intel, Microsoft, Adobe, Mersedes, BMW, etc. It is/was a solid brand, and more people wanted to have it, to be a part of it. People themselves raise the monsters, that eventually gobble everything by their insatiable greed.
And did it work out for them in the past?

The answer is no. Low price without feature parity does not increase market share. It will be seen as a poor man's alternative.
Not that i like AMD's current prices, but doing the same thing every time and expecting a different result is the definition of insanity. Plus almost complete absence in the OEM market where Nvidia cards go to prebulilt's.
This logic sounds, like: imagine yourself, someone, with his trousers completely sh*tted inside. Now that somone is unable to run or even walk. The smell is intolerable. But one still stands with face proud, making no look of failure.

nVidia, can afford to make any major scr*wup, and they have a "lagacy" of "reliability" and the image of being "best choice" for decades. Most media, as much as even the tech-savvy people, would rather give nVidia a benefit of the doubt, that tis is just "temporary" and not intentional.
Another issue, is that nVidia, has a long trace of buzzwordware, like GameWorks, HairWorks, PhysX, and many others, that "inflated" their GeForce brand like the baloon, for all these years. Thus, making it virtually more "valuable" than their "alternatives".
The GPP mentality is at work, and was since the inception of nVidia. Even in the days of first cards with GeForce monikers have appeared, and the nV VGAs rendering blurry smeary image, compared to the Radeon's, Rage, or any ohter rival. Somehow, everyone had their hair on fire from "green" brand products.

AMD on the other hand, cannot affort to do a mistake. Any, at all. They cannot do the same sh*t, as the monopoly player, and hope they will get away with it, as easy.
They cannot rely on being the -$50 alternaive (the end price is so inflated, that these fifty bucks difference is meaningless). Simply, because this doesn't work. Firstly, nVidia can drop the price (they obviously won't, ever), any moment. And secondly, being an underdog, one have suggest much more, or at least to price their products much more affordable, and attractive, to diversify the market. To the point, nobody would give a dang, about "extra" features, significantly more expencive GF products have. Instead, nobody gives a single eff, about AMD's products, because, they either priced similarly, or are non existant.
So people have to buy the available atrocious SKUs for the exorbitant prices. So, then the questioning is cutterly reasonable- why buy the product, that has less features, and preformance, if it costs same.

AMD could also make a ton of prebuilts, and entice the SI somehow, with their products. So far, AMD only showed, that they value their consumer GPU market as trash, and can abandon it any moment, for the sake of DC and Enterprise. Who gonna invest in that. Also the availability of AMD solutions, is an utter garbage. With pricing of their newest, and best, or even older stuff being a final nail into the coffin.

And nothing will change, since amd is already copying whatever nvidia is doing, them pulling out of the gpu market will have less than 0 impact. Just - they need to make sure they close the door on their way out, aight?
Exactly! Either they innovate, and suggest a better products. Or becime a second choice, due to the similar price.
Not surprising to see people still bought rtx 50 series cards despite all of the issues such as melting connectors, missing rops, and multiple driver updates causing black screens and crashes. The mindshare, marketing, and brand recognition are powerful enough to convince consumers to buy overpriced cards regardless of the problems.
And the gaming market is so screwed when people will empty their wallet for the leather jacket man no matter what, AMD cutting prices has never worked to outsell the competition.
It worked, and did quite well. AMD did sell tons of GPUs before. Price often can be a deciding factor. And in these days of harsh economy, this is especially actual. Problem is, AMD does't get this anymore. They won't get away any dollar, from their "premium brand".
And nobody follows the second place. Everyone is after the leader. Unless, the alternative attracts the good attention. If AMD treat themselves, as a second grade, a slightly cheaper underdog alternative, nothing will help.
What you talking about? I'm saying the 9070xt should be 25% cheaper than 5070ti than it being on par in cost. And AMD hasn't been doing what you say thry are. They been releasing rather lackluster gpus at near same coat as nvidia gpus. With the odd exceptions here and there like 6700xt as example.

Now mind you, I do agree with others saying that Nvidia has a cult following that really will buy anything Nvidia releases. Hell, we got member(s) who do just that. So I understand strength of brand loyalty too.
Again, $50 with current market situation, is naught. This basically is the same price, especially when thedifferent AIB's models come into play.
Nvidia launched new cards January 30th of Q1, AMD intentionally held cards in warehouses and retail shops until releasing them March 6 of Q1. The article is only referencing market share change over those 3 months, so this seems like the obvious outcome.
Still, the time won't help. This was obvious, that these delays would bite AMD in their greedy arses. The articles like this, is the best example. No matter how good the product is, but nobody will remember it, compared to the mistake associated with it's launch. ANd rivals would do everything, to make it spread as much as possible,
The 907XT should have been on sale, just after CES. ANd 9060XT, shuld have been on shelves, not later than couple weeks afte the former.
Maybe AMD should try making a better product next time.
Like it was mentione coutless times before, the products, especially RDNA 2 and 4 are fine. The AMD management, their dumb decisions, their delays, to price their SKUs accordingly to the nVidia's alternatives, is what did it disservice.
Right?! IDK why people are freaking out over these numbers. The 9070XT didn't even launch until March. Wait for Q2 to see the real head to head sales numbers.
A bit too late. Nobody will think about Q2 reports, after these "sensationalist's ones. This is the point. AMD had an atrocious launch. They have to do everything well, from the very beginning.
Nvidia is fudging the figures by including dGPUs meant for the AI market.
They did it countless of times. First, when the board have shat their pant's, when during "Pascal" mining crase, they could become liable, for substituting numbers in reports, for gamers cards sold as compute ones, for crypto. But then, the taste of money made them all feel calm. And after the AI bubble have began, nobody goves a single dang bout it.
 
Last edited:
Going Fabless works great if there is actually Fabs capable of handling the demand.

It is clear TSMC isn't up to the task. Maybe the companies need to start partnering up with other FAB's to try and do something. Since all the Lithography equipment come from ASML and not TSMC, they too could help invest in the fabrication process and modernized nodes like lets say Samsung or Toshiba(?). Look to get it produced elsewhere.
 
I love AMD I really do but they are the absolute KINGS of bad timing and missed opportunities, I mean sure this round while making a very capapble card for the sector it was as usual too late. The opportunity of game releases launching with FSR4 missed. it's almost a pattern. So these numbers come as no surprise, any possible gains were lost by not launching RDNA4 sooner before NV it may have upped their numbers a few points. One day AMD you will get your ducks in a row and maybe finally get an actual win rather than the constant game of catch up....
 
Going Fabless works great if there is actually Fabs capable of handling the demand.

It is clear TSMC isn't up to the task. Maybe the companies need to start partnering up with other FAB's to try and do something. Since all the Lithography equipment come from ASML and not TSMC, they too could help invest in the fabrication process and modernized nodes like lets say Samsung or Toshiba(?). Look to get it produced elsewhere.

What other fabs? The closest thing to TSMC's technologies right now would be Intel's 3 and 18A nodes, and the situation at Intel seems quite unclear at this point. There's no use hiding, Intel is struggling big time. Globalfoundries never developed anything past 12LP, and Samsung's latest nodes lack the transistor density of TSMC's - their very latest SF2 node that is supposed to enter production this year is roughly comparable to TSMC N3E.

It's not like TSMC is just hoarding fab equipment from ASML who are actually the ones responsible for all the tech, it doesn't work that way. If anything, ASML builds machines to TSMC's precise specification, not the other way around.
 
What other fabs? The closest thing to TSMC's technologies right now would be Intel's 3 and 18A nodes, and the situation at Intel seems quite unclear at this point. There's no use hiding, Intel is struggling big time. Globalfoundries never developed anything past 12LP, and Samsung's latest nodes lack the transistor density of TSMC's - their very latest SF2 node that is supposed to enter production this year is roughly comparable to TSMC N3E.

It's not like TSMC is just hoarding fab equipment from ASML who are actually the ones responsible for all the tech, it doesn't work that way. If anything, ASML builds machines to TSMC's precise specification, not the other way around.
If you bothered to read what I said, you just agreed with me.

Since all the Lithography equipment come from ASML

There is samsungs plant too which I said if maybe invest in to help get their nodes up to date to compete.

Would be a win win really, since it removes the monopoly that TSMC has, and also means more chips.
 
Last edited:
What other fabs? The closest thing to TSMC's technologies right now would be Intel's 3 and 18A nodes, and the situation at Intel seems quite unclear at this point. There's no use hiding, Intel is struggling big time. Globalfoundries never developed anything past 12LP, and Samsung's latest nodes lack the transistor density of TSMC's - their very latest SF2 node that is supposed to enter production this year is roughly comparable to TSMC N3E.

It's not like TSMC is just hoarding fab equipment from ASML who are actually the ones responsible for all the tech, it doesn't work that way. If anything, ASML builds machines to TSMC's precise specification, not the other way around.


I think there are some details about TSMC that goes beyond just the node size. Like all the new GPUs are on N4P, which is really an update of N5 and less dense then Intel 3 and Samsung's SF3E. Even Intel's own B580 is on some N5 derivative. Intel's node has been shipping for a year and Samsung's about 2 years. I have to conclude Intel and Samsung never put the investment in to get high production from these 3nm nodes.

18A is supposedly Intel's crown jewel and where much of the effort and $$ is at, so perhaps they'll see some traction there. I sometimes wonder though if both Intel and Samsung would not have been better off just focusing on and perfecting their 3nm nodes. After all, when we see the next round of GPUs in 18 months or so they are liable to be on N3 class nodes.
 
Or if it overheats and the connectors are unstable and melting.

But hey, let the ignorant and uninformed buy it though.

I'm just glad I could find mine @ MSRP.

MSRP ... I don't know what that is anymore. Saw an add for a car (won't name it so thread doesn't veer off over the selection) .. was $28,865. Called the dealer about the car in the ad. He said they have two. I asked, "How much if cash ?" He said $42,965 ... I said "I may be old but every time i bought a car in my life, I paid less than MSRP". His response was that the cars are so hard to get and keep in stock they have to charge that much to keep the doors open.

Meanwhile .... a PC builder in the US has to pay a 25% tariff on a video card, meanwhile the no name PCs you see in big box stores arrive in the US in the same boxes as they left China in but because they are 100% made in china from 100% chinese parts, they fall under an allowable exemption.
 
MSRP ... I don't know what that is anymore. Saw an add for a car (won't name it so thread doesn't veer off over the selection) .. was $28,865. Called the dealer about the car in the ad. He said they have two. I asked, "How much if cash ?" He said $42,965 ... I said "I may be old but every time i bought a car in my life, I paid less than MSRP". His response was that the cars are so hard to get and keep in stock they have to charge that much to keep the doors open.

Meanwhile .... a PC builder in the US has to pay a 25% tariff on a video card, meanwhile the no name PCs you see in big box stores arrive in the US in the same boxes as they left China in but because they are 100% made in china from 100% chinese parts, they fall under an allowable exemption.

There are no tariffs on computers, accelerators including GPUs, or smartphones.
 
Back
Top