john_AMD could similarly keep the RX 7000 and RX 6000 series in active production, provided they gave them native FSR 4 support right out of the gate. AMD's case might be slightly more complex since both old and new generations rely on standard GDDR6, but it could still make sense.

john_I think about this often and sometimes write about it as a potential solution. Yes, older nodes mean lower boost clocks and higher power consumption, but if that could alleviate market demand, why not? In the CPU space, partly thanks to cheaper DDR4, we see consumers actively choosing previous-generation processors, which fortunately remain available in decent quantities and at competitive prices.

Nvidia could easily resume production of the RTX 30 series. If they brought Frame Generation to that architecture—considering these chips would be manufactured at Samsung using an older yet mature process node and utilizing standard GDDR6/X—their price tags would unquestionably sit well below the RTX 50 series. Furthermore, because it's an older architecture, it likely wouldn't become a prime target for those hoarding hardware for AI workloads.

AMD could similarly keep the RX 7000 and RX 6000 series in active production, provided they gave them native FSR 4 support right out of the gate. AMD's case might be slightly more complex since both old and new generations rely on standard GDDR6, but it could still make sense.

The real hurdle is that engineering a new series to be manufactured across two entirely different fabrication nodes requires (hundreds of) millions in R&D investment. That said, if the rumors regarding Samsung pressuring AMD to offload part of its wafer production to Samsung's foundries hold true, we might just see a move like this in the near future.

BoggledBeagleI do not work for any manufacturer, but I believe I can view things objectivelly. And objectivelly I do not see many consumers holding back and not buying overpriced products.

BoggledBeagleLuxury good is an economic term, such products usually are ignored by state regulatory agencies and the pricing and purchasing is left on producers and consumers to figure out themselves alone. Consumers receive no state helping or protective hand. They are left to make their decisions for themselves.

Dr. DroNot feasible for several reasons, from using the same node as high-demand Ryzen and Epyc CPUs to their pretty much abandoned driver support (in the case of the 6000 series). I told all of you guys for years AMD were dips for omitting things like software DXR, alas, people with 5700 XTs can't even run games like Gears E-Day today. It's almost as if I wasn't just blindly hating and telling you guys that day would eventually come, and it sure as heck did

john_And we've told you plenty of things regarding Nvidia, too. The difference is that a lot of what you say about AMD we actually accept, whereas out of everything we're telling you about Nvidia, you refuse to accept a single thing.

Dr. DroNone of the doomposting or glum theories about how they are abandoning gamers came to pass, however.

john_So what we have is an Nvidia that originally launched the Titan lineup to push gamers toward accepting higher price points, eventually did the math, realized it was cannibalizing high-margin Quadro sales, and stripped away those workstation privileges from Titan GPUs. Then they reversed course briefly to quell pushback from buyers who had paid Titan money only to end up with glorified gaming cards, until they finally killed off the Titan brand altogether and absorbed the tier directly into the consumer gaming lineup as xx90 models.

Dragam1337It is quite clear that gamers are getting a shorter end of the stick for every day though :)

TesselatorAt this point, they're just fishing and it looks like it's fished out so they're reestablishing that segmentation once again. I wouldn't be surprised if the 6000 series didn't have a 90 at all but we saw a different naming convention emerge - though maybe not "Titan" again, I dunno.

kondaminpitty, I guess.
I wonder if those cards will still be worth anything when the next generation finally drops

Dr. DroNone of the doomposting or glum theories about how they are abandoning gamers came to pass, however. And yes, I realize the irony of saying this on a thread about "flagship gaming model is being discontinued", mostly because the whole situation is untenable to begin with.

YouseethisWell you COULD buy the rtx pro 6000 and game on it... it is very expensive... o_O

YouseethisWell you COULD buy the rtx pro 6000 and game on it... it is very expensive... o_O

Dr. DroNone of the doomposting or glum theories about how they are abandoning gamers came to pass, however. And yes, I realize the irony of saying this on a thread about "flagship gaming model is being discontinued", mostly because the whole situation is untenable to begin with.

kondaminpitty, I guess.
I wonder if those cards will still be worth anything when the next generation finally drops

john_Many of those who bought these cards early on, even above MSRP, are probably happy with the whole situation today because they see a hobby market turning into a lucrative investment. This joy won't last long, though, as future card generations will come at inflated prices, and I wouldn't be surprised if their raster performance per watt doesn't change much, but rather improves in AI—which will translate to better performance in DLSS 5 and primarily there.

As for the RTX 5090 changing categories, honestly, we are shooting ourselves in the foot with the things I'm reading in here. No, let Nvidia release an "RTX 5090 Workstation Edition" with pro drivers first, and then yes, I will accept that the chip has changed tier because of AI. As long as the card is marketed as a gaming product with gaming drivers, under no circumstances will I call it a professional GPU or an equivalent to a Titan. The Titan used to come with semi-pro drivers. Let Nvidia release at least semi-pro drivers exclusively for the RTX 5090, and then we can talk.

john_Honestly, when we as consumers are willing to accept compromises—even based on current market conditions—we are simply rolling out the red carpet for trillion-dollar corporations to give us less in the future while asking us for more money. I sit here reading comments saying, "The RTX 3090 and RTX 4090 were successors to the Titan". No, they weren't. They were gaming cards with gaming drivers, marketed as gaming products. Let's not adopt narratives that work against our own interests, regardless of how logical they might seem, because corporations rarely say, "Logic dictates we lower prices by 20%, so let's cut them by 20%", or "This feature is supported by previous-gen hardware, so let's offer it freely since we've already taken their money—which usually includes a premium anyway."

TesselatorI figured I wasn't quoting enough for that to be understood correctly. I'm saying there used to be segmentation (by marketing only) between Gaming and Studio cards. (sure one had more VRAM - like the the 1080ti had more than the 1080, which had more than the 1050, etc. But the segmentation was almost purely marketing. This was good because basic gamers knew to stay in the numbering scheme. Then NVidia manipulated the playground by calling their Titan line "X090" - which is a remarketing. Now with this announcement, they're (maybe) trying to reestablish that segmentation once again.

This isn't arguable, it is a VERY well known history everyone (including NVidia) understood at the time. The manipulation has come to confused a few people I see though.

Anyway, that was only a supporting point for the argument that we should be blaming ourselves for $7,000 video cards as much or more than NVidia. Personally I think it's about 90% us and only 10% them but YMMV. They only acted as any corporation would be expected to by following what the most lucrative segment were telegraphing. It was us gamers at the time 3090 and on, who were telegraphing, now it's the studio guys with AI wanting "workstations". And we're seeing NVidia following that school of delicious fishies just like any hungry whale would and should do. Our time to correct that is over. The XX90 cards NEVER actually belonged to the gaming segment. I see this as a good and proper correction!

Meanwhile, there are now a bunch of (extra?) people who paid Titan prices for Titan "prosumer" grade cards. Cool! It's a shame that the 80 and 70 cards got their prices dragged upward with that trend tho - to be sure.

halcyonGood, good. Bigger the bubble, bigger the crash. In the mean time, all the other computer non-memory/CPU components prices will have to come down too as sales are tanking. Once the AI capex bubble bursts, the discounts will be legendary. The faster we get there, better it is for almost everyone.

AusWolfPersonally, I don't care what a GPU is called. I only care about its price and performance. The 5090 has always been way above what many gamers would consider realistic and reasonable. Sure, to be classed a workstation card, you need workstation drivers, but Titan was never a workstation product, either, as far as I remember. It was a flagship gaming GPU with a different name for that sweet e-peen factor. The 5090 is no different in my eyes. A gaming card with the price and performance of a workstation product. A bit of a no man's land, imo (I know that many of you guys disagree, but let's not start a fight about opinions).


What red carpet? Whatever we spend on our PC upgrades is pocket money for a company like Nvidia. The true money is in AI and datacentre. We've crossed the point where even if the entire planet's gamer population stopped buying new GPUs, it wouldn't make a dent on Nvidia's income. They would just shift the whole production to AI. Whatever we call cards like the 5090 doesn't matter.

john_Even if the RTX 5090's price hits $100,000, I will personally still call it a gaming card.

QUANTUMPHYSICSI got my 5090, 64GB DDR5, 12TB SSD and a 285k BELOW MSRP during Black Friday 2025.

AAT2250

$4999

Both my cousins got the same computer.
Warranty extensions...

We can ride out the crisis.

Dragam1337Newer achitectures can be adapted to older nodes, even if it then wouldn't bring about as much as an improvement. So they wouldn't necesarrily need to keep producing old gens such as the 3000 series, they would simply need to adapt new achitectures to a different node than what the pro cards use.

As for it costing additional money in R&D, yes ofc - but considering that gaming does still bring in many billions each year in of itself, doing that so they could make products that gamers are actually excited for would be well worth it.

yfn_ratchetThat's the rub though, isn't it? Adapting a newer architecture with the newest feature sets to an older production node only provides the benefit of having the newest features—and even then, the additional hardware increases the die footprint and makes yields worse. You might end up with a product that isn't much cheaper after months of backporting it. It often makes more sense to simply continue production on last-gen products, like we're seeing with production of the 3060/GA106 being spun back up. At this rate it might be worth the money to instead port the GA106 die forwards onto TSMC 4N, or maybe a competing '4nm' node based on availability. Intel 4 Ampere chips would be particularly funny, I think. Or, you know, they could revive AD107 instead and start pumping out 4060s. Probably the cheaper and easier option.