NVIDIA GeForce RTX 4070 Priced at $600

Actually, that's exactly what should happen. Old tech should have a massive price decrease. Incoming parts should be faster and cheaper. "Cheaper" being a relative expression of stating that for the same level of cost as the previous generation buyers are getting a higher performing card. Meaning at a lower cost users should be able to buy something that was at minimum equal to a higher tier product.

An i5 today, destroys a Xeon from 2010. Prices go down, performance goes up.

What nVidia should be doing is giving their board partners rebates and releasing the new product stack at similar pricing to the old product stack. That's what every other market does and what nVidia used to also do. But as we know, they have become bigger dicks towards their board partners (eating all the margin without any recourse) and they are holding onto crypto pricing for as long as possible.
It should have a massive price decrease, but when many paid too much for their inventory? The warehouses and importers get like a 3-5% cut off the hardware and make their money from moving mass volume, taking that sort of loss would ruin them and the repercussions across the supply chain I don't even want to think about. Nvidia is dicks to them because they are the problem in more cases than not. They are dicks but they aren't assholes.
 
It should have a massive price decrease, but when many paid too much for their inventory? The warehouses and importers get like a 3-5% cut off the hardware and make their money from moving mass volume, taking that sort of loss would ruin them and the repercussions across the supply chain I don't even want to think about. Nvidia is dicks to them because they are the problem in more cases than not. They are dicks but they aren't assholes.
Despite all the garbage they do, they still recognize at some level that making their board partners go broke is bad for them.

However this is just as bad for the board partners. They have old stock that isn't moving and nVidia is getting all of their profit now by forcing them to have to buy more product that also isn't selling.

The exact same things will still have to happen if they need to adjust 4000 series pricing, which they do if they actually want to move any volume. Rebate now or later. Again, they're holding onto crypto pricing, screwing their board partners and screwing consumers. Most of us are content to wait. There simply isn't demand for this poor pricing. Waiting longer isn't going to change that.
 
I think it would have been okay for 4070 at $600. A bit pricey, but it does have decent performance. The regular 4070 rumored specs and its price tag looks underwhelming. More like a 4060ti at best.

I do wonder what a regular 4060 will look like. Probably will be $400, $450 actual retail price.

Historically, you paid a premium for the last extra bit of performance (e.g. 4090). The higher end card (4080) was about half of that price, and the higher mid range (4070) and low mid range (4060) were somewhat lower than that.

Now, you pay linearly all the way up the product stack. The 4070Ti is about half the performance of a 4090 and costs half as much. The best value is the high end, high margin, high priced card.
 
Last edited:
Despite all the garbage they do, they still recognize at some level that making their board partners go broke is bad for them.

However this is just as bad for the board partners. They have old stock that isn't moving and nVidia is getting all of their profit now by forcing them to have to buy more product that also isn't selling.

The exact same things will still have to happen if they need to adjust 4000 series pricing, which they do if they actually want to move any volume. Rebate now or later. Again, they're holding onto crypto pricing, screwing their board partners and screwing consumers. Most of us are content to wait.
Board partners already paid, Nvidia doesn’t get paid when the GPU is sold at retail, Nvidia gets paid when the AIB orders the GPU kit which contains the silicon, ram, and various specific components to the actual board.

AIB’s are holding that pricing because that’s what they already paid for, so any rebates are going to need to come from them asking Nvidia for a refund after the fact. The problem is the AIBs who were seeing their largest profits ever over ordered and Nvidia simply supplied that demand and priced according to their costs. The Nvidia margins didn’t change much during the pandemic but their volume doubled and costs on components went up accordingly so the costs on the kits went up to match.

Nvidia is holding MSRP crypto pricing to save the AIBs, the new silicon kits are being sold to them at lower costs because the demand and supply chain is vastly improved. The AIBs get to keep their covid margins and use that buffer to offset the losses they are going to have to take on the 3000 series parts when the drop becomes something that is required.

AIBs got greedy and they caused this problem, Nvidia just gets the criticism.

Edit:
It should be noted this is a repeat of what happened with the excessive 1000 series overstock. NVidia in that time did offer their AIBs rebates and bought back the overstock and it landed them an investor lawsuit that they lost for their troubles.
 
Last edited:
Board partners already paid, Nvidia doesn’t get paid when the GPU is sold at retail, Nvidia gets paid when the AIB orders the GPU kit which contains the silicon, ram, and various specific components to the actual board.
Yes. This is in essence what I said. NVidia gets paid first. Board partners are left holding the inventory bag.
AIB’s are holding that pricing because that’s what they already paid for, so any rebates are going to need to come from them asking Nvidia for a refund after the fact. The problem is the AIBs who were seeing their largest profits ever over ordered and Nvidia simply supplied that demand and priced according to their costs. The Nvidia margins didn’t change much during the pandemic but their volume doubled and costs on components went up accordingly so the costs on the kits went up to match.

Nvidia is holding MSRP crypto pricing to save the AIBs, the new silicon kits are being sold to them at lower costs because the demand and supply chain is vastly improved. The AIBs get to keep their covid margins and use that buffer to offset the losses they are going to have to take on the 3000 series parts when the drop becomes something that is required.

AIBs got greedy and they caused this problem, Nvidia just gets the criticism.
From everything that I’ve seen since the EVGA leak, this isn’t true.

NVidia determines maximum board price a given chip can have. They also determine how much they charge for their kits. Which also means they control exactly how much margins they get and exactly how much margins it’s possible for their board partners to get (by implementing a price ceiling).

So, I fundamentally disagree with your position.

https://www.tomshardware.com/news/e...arket-reportedly-citing-conflicts-with-nvidia

Both videos suggest that EVGA feels that Nvidia has stifled it, and suggest they don't find out details like MSRP, costs to buy GPU chips and more until Nvidia announces them, often publicly on stage. There is also talk of price ceilings as well as not being able to customize cards by changing the specs, like adding more memory.
 
Last edited:
Yes. This is in essence what I said. NVidia gets paid first. Board partners are left holding the inventory bag.

From everything that I’ve seen since the EVGA leak, this isn’t true.

NVidia determines maximum board price a given chip can have. They also determine how much they charge for their kits. Which also means they control exactly how much margins they get and exactly how much margins it’s possible for their board partners to get (by implementing a price ceiling).

So, I fundamentally disagree with your position.
The other AIB’s already stepped in and called EVGA out on that as mostly BS, they set a ceiling yes but that is standard. EVGAs exchange and warranty program was called out by Gigabyte and Asus as unsustainable and it it could be done they would have been offering the same program. There was a reason they were the only ones doing it, and them doing that is what cost them their margins which made them want better margins to cover for it.
 
The other AIB’s already stepped in and called EVGA out on that as mostly BS, they set a ceiling yes but that is standard. EVGAs exchange and warranty program was called out by Gigabyte and Asus as unsustainable and it it could be done they would have been offering the same program. There was a reason they were the only ones doing it, and them doing that is what cost them their margins which made them want better margins to cover for it.
So nVidia doesn’t implement a price ceiling, determine pricing of chips/boards, control modification of boards, or state the MSRP? Again: strongly disagree. All of these factors directly control nVidias margin and AIB margin. Feel free to post their “rebuttles” though. Will be interesting reading if nothing else.
 
Last edited:
So nVidia doesn’t implement a price ceiling, determine pricing of chips/boards, or state the MSRP? Again: strongly disagree. Feel free to post their “rebuttles” though. Will be interesting reading if nothing else.
They do implement a ceiling, but the GPU silicon is a very small part of the kits cost, Nvidia doesn’t get to choose the price of RAM, regulators, etc, they are just as much a victim of those costs as anybody else and those more than tripled in cost between 2020 and now.
As for the EVGA stuff I’m digging it up but here’s bits as I find them.
https://www.igorslab.de/en/evga-pulls-the-plug-with-loud-bang-yet-it-has-long-been-editorial/

The Igor’s lab summarizes it well enough but I can’t find the original sources they were on some Taiwanese sites I’d probably have better luck digging through the old threads here looking for them.
 
They do implement a ceiling, but the GPU silicon is a very small part of the kits cost, Nvidia doesn’t get to choose the price of RAM, regulators, etc, they are just as much a victim of those costs as anybody else and those more than tripled in cost between 2020 and now.
As for the EVGA stuff I’m digging it up but here’s bits as I find them.
https://www.igorslab.de/en/evga-pulls-the-plug-with-loud-bang-yet-it-has-long-been-editorial/

The Igor’s lab summarizes it well enough but I can’t find the original sources they were on some Taiwanese sites I’d probably have better luck digging through the old threads here looking for them.
Okay. Again to make this clear.
They can control the MSRP. the max price. And the margins they sell the kit for. We agree on all of these things. There is nothing you can state that tells me nVidia is powerless here.

Meanwhile there is tons of info stating how nVidia margin is increasing and AIB partner margin is decreasing.

I refer you to this handy chart:
C194F39C-1AEC-4DD1-9643-087B27A791CB.png

https://www.anandtech.com/show/17578/evga-and-nvidia-to-split-evga-wont-make-nextgen-nvidia-cards

And that comes from the original source, John Peddie Research here: https://www.jonpeddie.com/news/evga-wont-offer-nvidia-next-gen-series/

Yeah. No.

There is massive price inflation and it 100% has to do with nVidia slapping in more margin. It has very little to do with poor poor nVidia not getting to control RAM pricing. And all this has the AIB’s caught in the middle.
 
Last edited:
Okay. Again to make this clear.
They can control the MSRP. the max price. And the margins they sell the kit for. We agree on all of these things. There is nothing you can state that tells me nVidia is powerless here.

Meanwhile there is tons of info stating how nVidia margin is increasing and AIB partner margin is decreasing.

I refer you to this handy chart.
View attachment 561012
https://www.anandtech.com/show/17578/evga-and-nvidia-to-split-evga-wont-make-nextgen-nvidia-cards

Yeah. No.
That is specific to EVGA who gradually outsourced more and more while customizing more and more to an ever decreasing market as they only service North America and parts of Europe. EVGA’s margins were less than half of other AIBs as they kept far more in house and had a larger volume.

Regardless EVGA was circling the drain for years, they pulled the plug for a reason, they pidgin-holed themselves and their market was under heavy attack. Then there was the mass failure from New Worlds which Nvidia left them high and dry on because they were a result of EVGAs customization and not anything to do with the actual Nvidia silicon.
 
That is specific to EVGA who gradually outsourced more and more while customizing more and more to an ever decreasing market as they only service North America and parts of Europe. EVGA’s margins were less than half of other AIBs as they kept far more in house and had a larger volume.
You’ll note that chart says “AIB partners” and not “EVGA”. Reading the rest of the Peddie research article might help you.

Regardless EVGA was circling the drain for years, they pulled the plug for a reason, they pidgin-holed themselves and their market was under heavy attack. Then there was the mass failure from New Worlds which Nvidia left them high and dry on because they were a result of EVGAs customization and not anything to do with the actual Nvidia silicon.
They offered a better product, spent more in R&D and customization than any competitor while offering an industry leading RMA program, and weren’t allowed to price their product higher due to a price ceiling, a ceiling which you yourself recognize. They created a more expensive product and weren’t allowed to price it that way accordingly. Yes, leading to them having a much lower margin that their competitors.

EVGA did everything possible to have a product differentiated in design, cooling, overclocking, and software and were systematically not allowed to profit from their work. NVidia repeatedly worked either into silicon, contracts, or drivers stifling EVGA’s innovation including overclocking, increasing voltage limits, and proprietary cooling. Apparently we don’t deserve EVGA. Even with them having 40% of the North American market.

The only way you don’t see this is if you’re an nVidia shill. So whatever. I’m done, feel free to have the last word, but there is also no way you’re going to tell me that the 4080 “has to be” $1200 because poor poor nVidia can’t control component pricing. I’ve showed tons of evidence to the contrary.
 
Last edited:
You’ll note that chart says “AIB partners” and not “EVGA”.
And you will note in the article from Igors Lab, where the other AIB's say we make more than double that, and it was an EVGA problem.
They offered a better product, spent more in R&D and customization than any competitor while offering an industry leading RMA program, and weren’t allowed to price their product higher due to a price ceiling, a ceiling which you yourself recognize.
You can spend a crapload to make something better you reach a point where the cost doesn't meet the benefit, the best example of this with EVGA was their massive New Worlds failure, the replacement costs on the 3090s that their customizations caused to fail supposedly cost them most of their actual profits for that year.
EVGA did everything possible to have a product differentiated in design, cooling, overclocking, and software and were systematically not allowed to profit from their work. NVidia repeatedly worked either into silicon, contracts, or drivers stifling EVGA’s innovation including overclocking, increasing voltage limits, and proprietary cooling. Apparently we don’t deserve EVGA. Even with them having 40% of the North American market.
40% of the North American market is nothing, the Latin American market it 3x the size of the North American one, which is a fraction of the Asian market,, they tried to be different because they struggled to remain relevant, and now I worry about what I do should my EVGA 3080TI blow out, as my EVGA 2080TI did before it. (in all fairness the blowout was an EKWB failure, not an EVGA one so they got a pass). I mean I liked what EVGA tried to do but at some point, they being the squeaky wheel were bound to get the kick.

If you want to complain about Nvidia I am all for it, their licensing is atrocious, their pricing tiers punitive, and they are trying to force their smaller enterprise clients off their own hardware onto cloud services and into larger data centers to cut down on support costs by pricing the startup and upgrade costs to an unattainable point.

But actual hardware costs have a lot more complexity there aside from, GRRR Nvidia is being greedy!, and the AIBs share more blame for where we are right now than anybody else and I am tired of those bastards getting a pass like they didn't do this to all of us.
Nvidia cards are too expensive because Nvidia is greedy, meanwhile, those same AIBs jack their margins on Motherboards because they miss their revenue streams from when they were selling GPUs by the truckload to miners, and that's the only logical area they can make that shortfall.
I'm going for a drink, this is too depressing like I am watching my hobby/field collapse around me while people cheer for the cloud and the AI, there is so much blame to go around at this point it hurts to keep track of.
 
They offered a better product, spent more in R&D and customization than any competitor...

Not going to debate the other points as they are solid, but this seems like an EVGA problem. I only care about a few things:

1) Price
2) Performance (more in Nvidia's ballpark)
3) Cooling / noise
4) Simple OC software that can up the power limit, bump up the clocks, and adjust fan speed profiles if necessary

Most of the other brands offered similar things. I am not sure what EVGA's additional R&D and customization brought to the table for me. EVGA did have an ideal fan profile on the RTX 2070 series of coolers. It was very abrupt and would start on/off at idle during the hotter summer months. Their most recent OC software had a horrible GUI, made little logical sense, and often didn't save settings. Their old software from a decade ago was better.

My question is, what did this "customization" bring to the table that other brands did not do?
 
4) Simple OC software that can up the power limit, bump up the clocks, and adjust fan speed profiles if necessary
This is the crux of it, EVGA made its name on a warranty program that became increasingly unsustainable and its overclocking options.
As the cards grow more complex and the pressure from AMD increased the headroom in the silicon shrunk, just like the Intel and AMD fight on the CPU side, the AMD and Nvidia fight went the same way. Each side pushes the silicon as far as it can out of the box in a manner that makes sense. Any overclocking room left in the silicon that doesn't risk harm to the hardware itself is marginal at best, it mostly comes down to fan profiles and silent operation options now. If a vendor be it AMD, Intel, or Nvidia is comfortable enough to leave headroom for a meaningful overclock in their hardware it means that their competition has fallen flat and they are relaxed, there is no pressure to push things as far as they can. The ability to overclock a stock component without the need to resort to exotic cooling or power delivery methods means they intentionally left performance on the table because they were not pressured to push things further, we should not want that.
AMD and Nvidia have teams of engineers whose only job is to find a balance between price and performance for the silicon where things are viable for the market and I doubt EVGA and their comparatively minute team were capable of keeping pace and they knew it.

This was EVGAs whole model and it was being taken away from them, as a hobbyist, I miss being able to replace coolers and up voltage to grab an extra 15% on a card, but in hindsight that meant somebody working at either AMD or Nvidia failed at their job, they left that performance on the table not because they weren't smart enough to realize it was there, but because they didn't feel the need to reach for it.

Understand that the same people working in those AMD and Nvidia labs now, developing that hardware, those voltage and fan profiles, are those who wanted to work there, they started as hobbyists, and enthusiasts, people who spent their teens doing the same things we did, etching in pencil lines and swapping quartz crystals to get that little bit more.

In other news the bar I like has this new Squiffinator Dopplebock on tap and I can't figure out if I like it or not, it kicks like a horse, but it has this thick sweet toffee thing going, like coffee creamer almost and I somehow love and hate it all at once and I don't know how to proceed. Part of me says Poutine and the other part says Nachos.
 
This is the crux of it, EVGA made its name on a warranty program that became increasingly unsustainable and its overclocking options.
As the cards grow more complex and the pressure from AMD increased the headroom in the silicon shrunk, just like the Intel and AMD fight on the CPU side, the AMD and Nvidia fight went the same way. Each side pushes the silicon as far as it can out of the box in a manner that makes sense. Any overclocking room left in the silicon that doesn't risk harm to the hardware itself is marginal at best, it mostly comes down to fan profiles and silent operation options now. If a vendor be it AMD, Intel, or Nvidia is comfortable enough to leave headroom for a meaningful overclock in their hardware it means that their competition has fallen flat and they are relaxed, there is no pressure to push things as far as they can. The ability to overclock a stock component without the need to resort to exotic cooling or power delivery methods means they intentionally left performance on the table because they were not pressured to push things further, we should not want that.
AMD and Nvidia have teams of engineers whose only job is to find a balance between price and performance for the silicon where things are viable for the market and I doubt EVGA and their comparatively minute team were capable of keeping pace and they knew it.

This was EVGAs whole model and it was being taken away from them, as a hobbyist, I miss being able to replace coolers and up voltage to grab an extra 15% on a card, but in hindsight that meant somebody working at either AMD or Nvidia failed at their job, they left that performance on the table not because they weren't smart enough to realize it was there, but because they didn't feel the need to reach for it.

Understand that the same people working in those AMD and Nvidia labs now, developing that hardware, those voltage and fan profiles, are those who wanted to work there, they started as hobbyists, and enthusiasts, people who spent their teens doing the same things we did, etching in pencil lines and swapping quartz crystals to get that little bit more.

In other news the bar I like has this new Squiffinator Dopplebock on tap and I can't figure out if I like it or not, it kicks like a horse, but it has this thick sweet toffee thing going, like coffee creamer almost and I somehow love and hate it all at once and I don't know how to proceed. Part of me says Poutine and the other part says Nachos.

All brands have been offering slightly OCed models for so long now, but looking back at it there certainly were more brands that used to market this heavily. BFG, EVGA and if I recall XFX come to mind. For the past few GPU generations though it hasn't mattered as much. The last 4 or so GPUs I had would all clock higher than what was advertised on the box/website without even touching power limits. The last 3-4 GPUs I just turned up power limit (typically whatever the [H] benchmarks would recommend), and upped the core/memory clock around 85% of the way for what seemed to be the norm and left it. Always worked well.

Though I think EVGA managed to hold onto the market without relying on heavily over clocked models. Their main draw for me has been price. They were typically $30-40 cheaper in recent years than the competition.
 
That's a terrible comparison because anyone watching the car market also knows that there was a massive shortage there too due again to the silicon shortage. It was simply a function of being able to produce fewer cars, not because any of the components to make a car had gone up.

In fact Tesla has been lowering prices because the price of lithium has decreased (paywall sorry, but the point is it's true), something that no one thought would ever happen.

If you're paying close attention there is an expectation that the car market actually is going to implode. The silicon shortage is tapering off, so production has gone up. What has happened is that demand also has dropped like a rock. The ability of car manufacturers to command high prices for vehicles is ending. If you want to look into this and see evidence of that, look into what is happening with Carvana. They bought like crazy because they bet that skyrocketing prices would never end. Now demand has dropped in the used market at these inflated prices. They and everyone else will have to lower prices to get sales again. For them likely at an extreme loss.

This isn't inflation. It is a function of supply and demand. Supply side crashed due to the shortage so prices increased. Now supply side has come back but there isn't increased demand. Costs have to come down again to actually meet equilibrium. This is also directly what is happening in the GPU market. A majority of this is absolutely not inflation (some of it is, but not by 30+%). This is supply side trying to hang on to higher prices but failing hard because the demand isn't there at these prices. They have to reach equilibrium again for people outside of whales/enthusiasts to actually buy more (this is economics 101, supply and demand curves). We know with 100% certainty that we are not at economic equilibrium because video cards are sitting on shelves. Demand is not equal to supply. In fact: every publication has been stating that this is the lowest ever quarter in terms of CPU and GPU sales in over a decade.

If AMD/nVidia want GPU sales, then they're going to have to re-balance their curve and figure out how many GPU's they're going to need to sell to make similar profit as previous years. And they're going to have to take a margin cut or figure out cost cutting measures to make pricing in line with what people actually want to spend.
My point about vehicle costs still stands, you didn't negate it, you only looked at the past 5 years or so for your counter reply. Some companies benefit more due to a highly desired product that some people can afford, other companies are settled into a routine of things where there isn't much drastic change to their operation and prices are mostly stable.

Of course supply and demand factor in among other variables. However, when it comes to making chips, costs are increasing as the latest and greatest costs more and more to make smaller and smaller. Nvidia paid a premium to secure TSMC 3nm chip supply. Today you can render a movie like Toy Story on one GPU that we can go out and buy far faster than it took a fleet of computers 20 years ago, and it costs less than that fleet of computers did. People lose sight of where we came from. I am not trying to absolutely justify the price tags of current GPUs, but provide some context for why that price isn't really as bad as doomsday sayers are making it out to be. Who knows wht prices would look like without the crypto boom that blew up pricing. We need a correction from the fallout of that event. If people are not buying, Nvidia will have to lower prices to make more sales, but so many don't want to wait for that, they are dying for a new GPU now.
 
My point about vehicle costs still stands, you didn't negate it, you only looked at the past 5 years or so for your counter reply. Some companies benefit more due to a highly desired product that some people can afford, other companies are settled into a routine of things where there isn't much drastic change to their operation and prices are mostly stable.

Of course supply and demand factor in among other variables. However, when it comes to making chips, costs are increasing as the latest and greatest costs more and more to make smaller and smaller. Nvidia paid a premium to secure TSMC 3nm chip supply. Today you can render a movie like Toy Story on one GPU that we can go out and buy far faster than it took a fleet of computers 20 years ago, and it costs less than that fleet of computers did. People lose sight of where we came from. I am not trying to absolutely justify the price tags of current GPUs, but provide some context for why that price isn't really as bad as doomsday sayers are making it out to be. Who knows wht prices would look like without the crypto boom that blew up pricing. We need a correction from the fallout of that event. If people are not buying, Nvidia will have to lower prices to make more sales, but so many don't want to wait for that, they are dying for a new GPU now.
They're not going to lower prices, I think they're fine selling fewer cards and making higher margins. If they can sell fewer cards and be as profitable in the gaming market, why would they? I'm guessing you'll never see an official price cut on the 4080, but they'll drop 200 dollars on Newegg as a "deal." People either need to accept this is the new GPU market or buy a console. If I quit making games for a living I'll probably switch back to console gaming at this point.

TSMC 3nm is insanely expensive, I'd expect the 5080 to be 13-1400 when Blackwell launches. Wouldn't shock me if it costs the same as the 4090 today and the 5090 is 2000+.
 
They're not going to lower prices, I think they're fine selling fewer cards and making higher margins. If they can sell fewer cards and be as profitable in the gaming market, why would they? I'm guessing you'll never see an official price cut on the 4080, but they'll drop 200 dollars on Newegg as a "deal." People either need to accept this is the new GPU market or buy a console. If I quit making games for a living I'll probably switch back to console gaming at this point.

TSMC 3nm is insanely expensive, I'd expect the 5080 to be 13-1400 when Blackwell launches. Wouldn't shock me if it costs the same as the 4090 today and the 5090 is 2000+.
TSMC 3nm also has yield issues. That will certainly play a part if said issues are still plaguing them by the time 50-series starts production.

As for selling fewer cards in desktop, I agree. Nvidia seems content to switch from crypto miners to AI as their new cash cow.
 
That's a terrible comparison because anyone watching the car market also knows that there was a massive shortage there too due again to the silicon shortage. It was simply a function of being able to produce fewer cars, not because any of the components to make a car had gone up.

In fact Tesla has been lowering prices because the price of lithium has decreased (paywall sorry, but the point is it's true), something that no one thought would ever happen.

If you're paying close attention there is an expectation that the car market actually is going to implode. The silicon shortage is tapering off, so production has gone up. What has happened is that demand also has dropped like a rock. The ability of car manufacturers to command high prices for vehicles is ending. If you want to look into this and see evidence of that, look into what is happening with Carvana. They bought like crazy because they bet that skyrocketing prices would never end. Now demand has dropped in the used market at these inflated prices. They and everyone else will have to lower prices to get sales again. For them likely at an extreme loss.

This isn't inflation. It is a function of supply and demand. Supply side crashed due to the shortage so prices increased. Now supply side has come back but there isn't increased demand. Costs have to come down again to actually meet equilibrium. This is also directly what is happening in the GPU market. A majority of this is absolutely not inflation (some of it is, but not by 30+%). This is supply side trying to hang on to higher prices but failing hard because the demand isn't there at these prices. They have to reach equilibrium again for people outside of whales/enthusiasts to actually buy more (this is economics 101, supply and demand curves). We know with 100% certainty that we are not at economic equilibrium because video cards are sitting on shelves. Demand is not equal to supply. In fact: every publication has been stating that this is the lowest ever quarter in terms of CPU and GPU sales in over a decade.

If AMD/nVidia want GPU sales, then they're going to have to re-balance their curve and figure out how many GPU's they're going to need to sell to make similar profit as previous years. And they're going to have to take a margin cut or figure out cost cutting measures to make pricing in line with what people actually want to spend.
Automakers, at least Chevy and Ford decided that they can make a higher margin on each vehicle while keeping production low. For a while, it was out of necessity due to supply chain shortage. Now, they are content to keep production lower, shutter some production and lay off employees. It will take a few years for Toyota, Honda, etc. to take up the slack.


20230403_073523.jpg
 
Automakers, at least Chevy and Ford decided that they can make a higher margin on each vehicle while keeping production low. For a while, it was out of necessity due to supply chain shortage. Now, they are content to keep production lower, shutter some production and lay off employees. It will take a few years for Toyota, Honda, etc. to take up the slack.
Yeah pretty much companies got addicted to "pandemic" margins and being able to blame covid for literally every failing. Thanks people who bought in and supported this market. Whenever someone says "Why do you care how somebody spends their money?", remember this.
 
Automakers, at least Chevy and Ford decided that they can make a higher margin on each vehicle while keeping production low. For a while, it was out of necessity due to supply chain shortage. Now, they are content to keep production lower, shutter some production and lay off employees. It will take a few years for Toyota, Honda, etc. to take up the slack.
Are chevy-ford car for the USA market specially made more in the USA than Toyota-Honda car sold in the US market ?
 
Yeah pretty much companies got addicted to "pandemic" margins and being able to blame covid for literally every failing. Thanks people who bought in and supported this market. Whenever someone says "Why do you care how somebody spends their money?", remember this.
Yep. Done with the "COVID" excuses. These companies got addicted to the new margins and refuse to change. We as consumer have a choice not to support that.

And before someone asks, I got my card secondhand last year.
 
Kind

Sort of like how some 4070 Ti models are roughly the same price as a 4080.

A few of them in stock at around $830 to 850 from MSI, PNY and ASUS. Which is still very over priced, and I doubt $600 4070s will be all that common. A lot more around $880 though. Yes there are some that are $1000, but I am not sure who is buying those. ASUS has a model for $850 on newegg as well as $1050. Never seen $200 variance for the same type of card before.

Cheapest RTX 4080 I see is $1190 by MSI.

I am thinking the 4070 will land around $650 for basic models. Hopefully they at least have something for $600.
 
TSMC 3nm also has yield issues. That will certainly play a part if said issues are still plaguing them by the time 50-series starts production.

As for selling fewer cards in desktop, I agree. Nvidia seems content to switch from crypto miners to AI as their new cash cow.
Yeah, they are breaking the N3 process as it currently exists into a pair of new processes N3E and N3B.
N3B is the new renaming of the existing N3 node (which they need to fix), and TSMC has problems there it is failing to meet power, performance, or yield metrics
N3E is a relaxed version of N3, N3E uses single-pattern EUV instead of a multi-pattern process (https://semiengineering.com/single-vs-multi-patterning-euv/), it also changes the metal pitch and cuts back on layers going from 25 down to 19, so it results in a 1.3x gain over N5 instead of the initial 1.6x gain in density, but it does a better job on delivering power and performance metrics while vastly improving the failure rate. N3E uses the same cell size as N5 which is a very big deal for area scaling and SRAM if you want to know more there are lots of articles out there on that topic, but it is a lot to unpack. But unless TSMC can fix that issue the scaling advantage they have held over Intel is going to very quickly vanish.
 
Are chevy-ford car for the USA market specially made more in the USA than Toyota-Honda car sold in the US market ?
No idea, but I did a paper in 2000 for school on Honda in America. At that time Honda exported more cars out of America than any other company. I don't think they made more, but they exported more. Currently, it wouldn't surprise me if Honda and Toyota made more vehicles in the USA than Ford and Chevy, but I can't back it up.
 
Currently, it wouldn't surprise me if Honda and Toyota made more vehicles in the USA than Ford and Chevy, but I can't back it up.
Which is why I was not sure that I followed the link that cars made in the US was specially a reflexion about Ford and GM like in the sentence Automakers, at least Chevy and Ford decided that they can make a higher margin on each vehicle while keeping production low.

And the graph do seem to show an going to almost 10 years trend with not really much change, it could be true that GM-Ford went more into heavier vehicle (mustang aside, pickup-suv) with less overall sales strategy, GM did simplify a lot the line up and amount of brand over time
 
I’m hesitant to be saying a 1660 Super is still perfectly serviceable. Some of these games this year have been stupid in hardware requirements. It’s almost like they have to brute force bad coding. VRAM is going to also be what kills the sub 800 brackets.
 
I’m hesitant to be saying a 1660 Super is still perfectly serviceable. Some of these games this year have been stupid in hardware requirements. It’s almost like they have to brute force bad coding. VRAM is going to also be what kills the sub 800 brackets.
Nothing says you need to play all the triple A games that come out too. And FYI, I found that Far Cry 6 was perfectly playable on a gtx 970, there were some graphical glitches but nothing that made the game seem sluggish. Just need to have lower expectations on what level of graphics you're ok with, and yeah "low" is fine when the majority of playing is you moving around constantly with a focus more on your HUD and middle of the screen.
 
Back
Top