Nope. AMD was a CPU manufacturer that bought ATI, a GPU manufacturer to become the #2 GPU vendor.
NVIDIA was a GPU manufacturer that bought a license for the ARM architecture and developed their own CPU/GPU SOC that they demoed in the "NVIDIA Shield" player that later made a design win in the Nintendo Switch. NVIDIA tried to buy ARM but got blocked by regulators.
I have two guesses about what is going on.
One is that the benefits of smaller transistor sizes are getting reduced in generation after generation. From the 1980s to 2005 the clock rate would go up so you not only had more complex chips, but they got faster explosively. Post 2005 clock rates have been stagnant but they would give you more cores, better SIMD facilities and better instruction per clock.
Still the price per transistor would fall but we are on the cusp of the price not dropping when the die shrinks. The move to chiplets, advanced packaging, and all that doesn't really help because you are still assembling packages out of expensive chips and spending more money on additional manufacturing steps. (Even if it does mean you can build SKUs that wouldn't otherwise be practical by putting together parts from a library.)
(The worst thing on the horizon may be that the lifetime of transistors is going to get shorter in which case a 50-series will not only have an eye watering price but it will burn up in two years.)
All of that is pointing to GPU economics getting worse whether or not the vendors are profiteering.
You might say they'd have a reason not to profiteer to not hurt their business in the long term but I'm not sure what business they will have in the long term. PC and console gaming have been shrinking relative to mobile gaming in recent years
Mobile gaming is nearly half of the business now and growing, the whole industry is being remade in the shape of phones and maybe they figure the gaming PC you build now (if you build a gaming PC) will be the last one you ever build and they'd better take you for all you're worth.
(The cloud changes the economics too because potentially a $2000 GPU in the cloud can be utilized close to 100% by different users whereas the GPU in your PC will usually be used just when you're playing. By that standard NVIDIA and AMD can't afford to sell you a GPU when a cloud compute vendor could afford to pay more for it.)
I got really depressed about the state of the game industry when I went shopping for games that expired on my XBOX ONE and found that, unlike lists for Nintendo platforms, "top 25" and such lists for the XBOX ONE are full of titles I made up my mind about years ago like "Call of Duty", "Assassin's Creed" and "Madden NFL". Shockingly, Skyrim was on the list despite it being a bargain bin game when I had an Xbox 360. Unlike Japanese and Chinese game studios, Western companies are not introducing major new games, just making more "Assassin's Creed" games, the one new game I can think of is "Control" and that comes from Finland.
> PC and console gaming have been shrinking relative to mobile gaming in recent years
Is mobile gaming actually taking customers from PC though or is it just growing faster because there are a lot of accessible casual games? I know a lot of PC gamers and I can't imagine any of them swapping to gaming primarily on mobile. I don't know any people who are hobbyist mobile gamers.
That's how you and I think and it is how the Finnish and Japanese game makers think but it is not how Activision, Ubisoft, and EA think. At this point they've decided to disinvest in their legacy games. They are are intensely envious of the success of Genshin Impact which achieves AAA graphics and gameplay on mobile but they can't replicate its success because they are too afraid to take risks.
Despite the fact that mobile phones (ex. Apple) have not been improving since 2017 there is a perception of inevitability in the industry that phones will eat absolutely everything. They figure game addicts will switch to mobile games when their old hardware breaks down and they don't have a choice. (... they'll keep playing the same old games, the main thing you need is a decent game controller for mobile, there has to be an interesting story of why every company other than a big console manufacturer struggles to make game controllers that work -- look at the failures of Logitech, Steam, Google, etc.)
which may have been a bit too influential. He made the case that firms in industry after industry went under because they failed to invest in new technology because the new technology was not, at first, good enough to satisfy their legacy customers.
What we see now (Microsoft releasing a tablet OS for PCs in Windows 8, Facebook neglecting a very good monopoly business for dreams of virtual reality, ...) is firms frequently throwing their current customers under the bus in the name of what might be... Because they read that book and don't want to wind up like one of the case studies.
As it is now it is a form of corporate suicide and we are waiting for it to play out and have a B-school professor write a book about it. But when you look at this way many of the things that make no sense in gaming like GAME PASS and the many me too game streaming services are part of the strange spectacle of a game industry that is failing despite still-rising revenues. It's hard, for instance, to picture Nintendo coming out with a real sequel to the Switch for many reasons (e.g. they could make a console 4x as powerful... that costs 4x as much; who needs better platforms when you are just rehashing Diablo for the n-th time and CS:GO is still a leading game?)
It's enough to make me think I should quit playing games, go out for walk or pet a puppy or make some friends! Maybe more people will the way the industry is heading.
If I was going to get a handheld it would almost certainly be a Switch because of the strong collection. I do like Fire Emblem but the Switch has all the big Western game and all the Japanese games that used to come out for the Playstation Vita. The Deck very much follows in the footsteps of the Vita and most of those Japanese games are on Steam so I could have a lot of fun w/ a Deck but (1) I just built a gaming desktop that plays Steam games, and (2) my handheld is still a New Nintendo 3DS with a backlog.
Yeah I think you're definitely right wrt the attitude big studios are taking with their games these days. If they swap to mobile they might pick up a different audience but someone will have to fill the void with quality PC games. If Finnish and Japanese (and small US) studios fill that void I'm totally A-OK with that personally.
> I should quit playing games
Sadly I'm right there with you... life has gotten too busy and most of my old gamer friends have moved on to other games that would require a huge time investment to get up to speed in. Now I just play casual games with IRL friends and my brother, haven't played a PC game in ages. Fingers crossed for Starfield though.
I wouldn't expect any company to continue running their factories at peak capacity when there is weak demand. The outcome of doing so is, of course, destroying your own profit. And while that sounds great in theory for consumers (cheaper prices!), in a capital intensive industry, this is a great way to go out of business.
Another example from another industry: agriculture subsidies cause farmers to do exactly this, producing massive amounts of crops despite a low price for them; a wasteful outcome by most societal standards, and the low profit margin for farming means none but the most efficient, massive operations can survive.
In essence, demand from crypto that ended up vanishing to a great extent overnight (ETH Merge), supply chain disruptions and shaky economy means supply and demand is less and less predictable for these companies.
It makes absolute sense to try and keep it stable. Price stickiness aside, building chips is a long term commitment, if they cannot somehow stabilise it, their investments into future chips will be riskier as well (read: more expensive, to compensate for it).
17 comments
[ 2.9 ms ] story [ 46.2 ms ] threadNVIDIA was a GPU manufacturer that bought a license for the ARM architecture and developed their own CPU/GPU SOC that they demoed in the "NVIDIA Shield" player that later made a design win in the Nintendo Switch. NVIDIA tried to buy ARM but got blocked by regulators.
I have two guesses about what is going on.
One is that the benefits of smaller transistor sizes are getting reduced in generation after generation. From the 1980s to 2005 the clock rate would go up so you not only had more complex chips, but they got faster explosively. Post 2005 clock rates have been stagnant but they would give you more cores, better SIMD facilities and better instruction per clock.
Still the price per transistor would fall but we are on the cusp of the price not dropping when the die shrinks. The move to chiplets, advanced packaging, and all that doesn't really help because you are still assembling packages out of expensive chips and spending more money on additional manufacturing steps. (Even if it does mean you can build SKUs that wouldn't otherwise be practical by putting together parts from a library.)
(The worst thing on the horizon may be that the lifetime of transistors is going to get shorter in which case a 50-series will not only have an eye watering price but it will burn up in two years.)
All of that is pointing to GPU economics getting worse whether or not the vendors are profiteering.
You might say they'd have a reason not to profiteer to not hurt their business in the long term but I'm not sure what business they will have in the long term. PC and console gaming have been shrinking relative to mobile gaming in recent years
https://kevurugames.com/blog/mobile-gaming-vs-pc-gaming-over...
Mobile gaming is nearly half of the business now and growing, the whole industry is being remade in the shape of phones and maybe they figure the gaming PC you build now (if you build a gaming PC) will be the last one you ever build and they'd better take you for all you're worth.
(The cloud changes the economics too because potentially a $2000 GPU in the cloud can be utilized close to 100% by different users whereas the GPU in your PC will usually be used just when you're playing. By that standard NVIDIA and AMD can't afford to sell you a GPU when a cloud compute vendor could afford to pay more for it.)
I got really depressed about the state of the game industry when I went shopping for games that expired on my XBOX ONE and found that, unlike lists for Nintendo platforms, "top 25" and such lists for the XBOX ONE are full of titles I made up my mind about years ago like "Call of Duty", "Assassin's Creed" and "Madden NFL". Shockingly, Skyrim was on the list despite it being a bargain bin game when I had an Xbox 360. Unlike Japanese and Chinese game studios, Western companies are not introducing major new games, just making more "Assassin's Creed" games, the one new game I can think of is "Control" and that comes from Finland.
Is mobile gaming actually taking customers from PC though or is it just growing faster because there are a lot of accessible casual games? I know a lot of PC gamers and I can't imagine any of them swapping to gaming primarily on mobile. I don't know any people who are hobbyist mobile gamers.
Despite the fact that mobile phones (ex. Apple) have not been improving since 2017 there is a perception of inevitability in the industry that phones will eat absolutely everything. They figure game addicts will switch to mobile games when their old hardware breaks down and they don't have a choice. (... they'll keep playing the same old games, the main thing you need is a decent game controller for mobile, there has to be an interesting story of why every company other than a big console manufacturer struggles to make game controllers that work -- look at the failures of Logitech, Steam, Google, etc.)
Clayton Christensen wrote this influential book
https://www.amazon.com/Innovators-Dilemma-Revolutionary-Chan...
which may have been a bit too influential. He made the case that firms in industry after industry went under because they failed to invest in new technology because the new technology was not, at first, good enough to satisfy their legacy customers.
What we see now (Microsoft releasing a tablet OS for PCs in Windows 8, Facebook neglecting a very good monopoly business for dreams of virtual reality, ...) is firms frequently throwing their current customers under the bus in the name of what might be... Because they read that book and don't want to wind up like one of the case studies.
As it is now it is a form of corporate suicide and we are waiting for it to play out and have a B-school professor write a book about it. But when you look at this way many of the things that make no sense in gaming like GAME PASS and the many me too game streaming services are part of the strange spectacle of a game industry that is failing despite still-rising revenues. It's hard, for instance, to picture Nintendo coming out with a real sequel to the Switch for many reasons (e.g. they could make a console 4x as powerful... that costs 4x as much; who needs better platforms when you are just rehashing Diablo for the n-th time and CS:GO is still a leading game?)
It's enough to make me think I should quit playing games, go out for walk or pet a puppy or make some friends! Maybe more people will the way the industry is heading.
If I was going to get a handheld it would almost certainly be a Switch because of the strong collection. I do like Fire Emblem but the Switch has all the big Western game and all the Japanese games that used to come out for the Playstation Vita. The Deck very much follows in the footsteps of the Vita and most of those Japanese games are on Steam so I could have a lot of fun w/ a Deck but (1) I just built a gaming desktop that plays Steam games, and (2) my handheld is still a New Nintendo 3DS with a backlog.
> I should quit playing games
Sadly I'm right there with you... life has gotten too busy and most of my old gamer friends have moved on to other games that would require a huge time investment to get up to speed in. Now I just play casual games with IRL friends and my brother, haven't played a PC game in ages. Fingers crossed for Starfield though.
https://seekingalpha.com/article/4573788-intel-q4-earnings-b...
If you dislike this behavior, start a chip company and undercut the market!
Another example from another industry: agriculture subsidies cause farmers to do exactly this, producing massive amounts of crops despite a low price for them; a wasteful outcome by most societal standards, and the low profit margin for farming means none but the most efficient, massive operations can survive.
In essence, demand from crypto that ended up vanishing to a great extent overnight (ETH Merge), supply chain disruptions and shaky economy means supply and demand is less and less predictable for these companies.
It makes absolute sense to try and keep it stable. Price stickiness aside, building chips is a long term commitment, if they cannot somehow stabilise it, their investments into future chips will be riskier as well (read: more expensive, to compensate for it).
What happens is when prices rise, third party investment rises because of opportunity for profit is easier.
This is what Intel did for years and is now suffering... they'd been better off shipping at rock bottom prices and locking out competition.
I can't wait till we have something like the Nvidia Grace processor available from SuperMicro. Hoping to kiss some of our x86 servers goodbye!