6 ms·

I don't think Nvidia is worried about cannibalizing their A100 and H100 business, I think they're trying to keep the cards something resembling affordable for g

I don't think Nvidia is worried about cannibalizing their A100 and H100 business, I think they're trying to keep the cards something resembling affordable for gamers.

AnthonyMouse 3y ago
Then why wouldn't they just make more of them? The excuse is supposed to be fab capacity, but the 3070 has better performance than the 4060 etc. and is built on the older process which should no longer be in short supply.
- autoexec 3y ago
  Is it an issue of supply? I figure they think non-gamers will be willing and able to pay far more for their hardware than most gamers could ever afford. If they sold comparable products to gamers at a low price the non-gamers would just end up buying them up and even if there was enough supply to satisfy the needs of everyone Nvidia would still be out a fortune. It'd be better for them to sell low priced gaming cards that would perform poorly for non-gaming purposes and sell extremely high priced specialty cards to the people who want to use them for AI or crypto or whatever other non-gaming uses they come up with. That'd at least keep the price of video cards low for gamers, avoid supply issues, and allow nvidia to extract massive amounts of profit from companies with no interest in video games. The only downside would be that it makes it harder for anyone who doesn't have deep pockets to get into the AI game.
  - AnthonyMouse 3y ago
    > If they sold comparable products to gamers at a low price the non-gamers would just end up buying them up and even if there was enough supply to satisfy the needs of everyone Nvidia would still be out a fortune. That's the point. They're not trying to do gamers a favor, if it was that they'd just make more cards. What they're trying to do is market segmentation, which customers despise and resent.
    - creato 3y ago
      I don’t know, in the world you want, gamers don’t have GPUs, and NVIDIA has less money. The only winners are people buying hardware for data centers.
      - AnthonyMouse 3y ago
        They could leave the connector on the GPUs and then make more of them. Then everybody has more GPUs and everybody wins except Nvidia. Or to put it the other way, in the status quo everybody loses except Nvidia.
  - solardev 3y ago
    They used to do this with the Quattro workstation cards. What happened to those?
- aurareturn 3y ago
  Then why wouldn't they just make more of them? The excuse is supposed to be fab capacity, but the 3070 has better performance than the 4060 etc. and is built on the older process which should no longer be in short supply. There is actually very little margins in the midrange consumer discrete GPU market. The market for discrete GPUs have been shrinking since the mid 2000s.[0] Most GPUs are sold as integrated such as SoCs and in consoles nowadays. In a shrinking market, the midrange and low-end products will cease to be profitable. Hence, Nvidia's 60/70 offerings are lackluster because they don't make much money from them. They want you to buy the 80s and 90s cards. Furthermore, node advancements have stopped scaling $/transistor. So the transistors aren't getting cheaper, just smaller. Lastly, Nvidia wants to allocate every last wafer they pre-purchased from TSMC to their server GPUs. [0]https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2022/12/GPU_Dev_graph_01.png https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1...
  - AnthonyMouse 3y ago
    > Most GPUs are sold as integrated such as SoCs and in consoles nowadays. But those are mostly AMD, and doesn't really have anything to do with what features someone puts on their discrete gaming cards, except insofar as it implies gamers don't need the cards some amateur ML hobbyist might buy. > In a shrinking market, the midrange and low-end products will cease to be profitable. That's assuming the products have high independent development costs, but that isn't really the case. The low end products are essentially the high end products with fewer cores which use correspondingly less silicon -- which have higher yields because you don't need such a large area of perfect silicon or can sell a defective die as a slower part by disabling the defective section, making them profitable with a smaller margin per unit die area. > Furthermore, node advancements have stopped scaling $/transistor. So the transistors aren't getting cheaper, just smaller. Which implies that they can profitably continue producing almost-as-good GPUs on the older process node. > Lastly, Nvidia wants to allocate every last wafer they pre-purchased from TSMC to their server GPUs. Which is why the proposal is for them to make as many GPUs as the gamers could want at Samsung.
    - aurareturn 3y ago
- ksec 3y ago
  >and is built on the older process which should no longer be in short supply. I am not sure where that idea came from. It is from Samsung 8nm Fab. It never had the capacity to play with in the first place. Especially when Samsung Foundry is upgrading to chase with leading node.
  - AnthonyMouse 3y ago
    In general you want to upgrade your oldest fab to the newest node, or build a new one, instead of shutting down a recent one and taking it out of production even though there is still demand for that node. There are still plenty of things being produced in fabs with older technology than that. Global Foundaries is the third largest in the world and they're offering 12nm or worse. People buy it because not everything needs a node which less than six months old and the price is right.
    - ksec 3y ago
      Yes. But that is specific to GF and TSMC when you have recurring customer on older node. Which is not true for Samsung Foundry.
      - AnthonyMouse 3y ago
        Why is it different? Wouldn't Samsung prefer to build new fabs rather than retrofitting old ones given the demand, since they would then have more customers and make more money?

I don't think Nvidia is worried about cannibalizing their A100 and H100 business, I think they're trying to keep the cards something resembling affordable for gamers.

AnthonyMouse 3y ago

Then why wouldn't they just make more of them? The excuse is supposed to be fab capacity, but the 3070 has better performance than the 4060 etc. and is built on the older process which should no longer be in short supply.

autoexec 3y ago
Is it an issue of supply? I figure they think non-gamers will be willing and able to pay far more for their hardware than most gamers could ever afford. If they sold comparable products to gamers at a low price the non-gamers would just end up buying them up and even if there was enough supply to satisfy the needs of everyone Nvidia would still be out a fortune. It'd be better for them to sell low priced gaming cards that would perform poorly for non-gaming purposes and sell extremely high priced specialty cards to the people who want to use them for AI or crypto or whatever other non-gaming uses they come up with. That'd at least keep the price of video cards low for gamers, avoid supply issues, and allow nvidia to extract massive amounts of profit from companies with no interest in video games. The only downside would be that it makes it harder for anyone who doesn't have deep pockets to get into the AI game.
- AnthonyMouse 3y ago
  > If they sold comparable products to gamers at a low price the non-gamers would just end up buying them up and even if there was enough supply to satisfy the needs of everyone Nvidia would still be out a fortune. That's the point. They're not trying to do gamers a favor, if it was that they'd just make more cards. What they're trying to do is market segmentation, which customers despise and resent.
  - creato 3y ago
    I don’t know, in the world you want, gamers don’t have GPUs, and NVIDIA has less money. The only winners are people buying hardware for data centers.
    - AnthonyMouse 3y ago
      They could leave the connector on the GPUs and then make more of them. Then everybody has more GPUs and everybody wins except Nvidia. Or to put it the other way, in the status quo everybody loses except Nvidia.
- solardev 3y ago
  They used to do this with the Quattro workstation cards. What happened to those?
aurareturn 3y ago
Then why wouldn't they just make more of them? The excuse is supposed to be fab capacity, but the 3070 has better performance than the 4060 etc. and is built on the older process which should no longer be in short supply. There is actually very little margins in the midrange consumer discrete GPU market. The market for discrete GPUs have been shrinking since the mid 2000s.[0] Most GPUs are sold as integrated such as SoCs and in consoles nowadays. In a shrinking market, the midrange and low-end products will cease to be profitable. Hence, Nvidia's 60/70 offerings are lackluster because they don't make much money from them. They want you to buy the 80s and 90s cards. Furthermore, node advancements have stopped scaling $/transistor. So the transistors aren't getting cheaper, just smaller. Lastly, Nvidia wants to allocate every last wafer they pre-purchased from TSMC to their server GPUs. [0]https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2022/12/GPU_Dev_graph_01.png https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1...
- AnthonyMouse 3y ago
  > Most GPUs are sold as integrated such as SoCs and in consoles nowadays. But those are mostly AMD, and doesn't really have anything to do with what features someone puts on their discrete gaming cards, except insofar as it implies gamers don't need the cards some amateur ML hobbyist might buy. > In a shrinking market, the midrange and low-end products will cease to be profitable. That's assuming the products have high independent development costs, but that isn't really the case. The low end products are essentially the high end products with fewer cores which use correspondingly less silicon -- which have higher yields because you don't need such a large area of perfect silicon or can sell a defective die as a slower part by disabling the defective section, making them profitable with a smaller margin per unit die area. > Furthermore, node advancements have stopped scaling $/transistor. So the transistors aren't getting cheaper, just smaller. Which implies that they can profitably continue producing almost-as-good GPUs on the older process node. > Lastly, Nvidia wants to allocate every last wafer they pre-purchased from TSMC to their server GPUs. Which is why the proposal is for them to make as many GPUs as the gamers could want at Samsung.
  - aurareturn 3y ago
ksec 3y ago
>and is built on the older process which should no longer be in short supply. I am not sure where that idea came from. It is from Samsung 8nm Fab. It never had the capacity to play with in the first place. Especially when Samsung Foundry is upgrading to chase with leading node.
- AnthonyMouse 3y ago
  In general you want to upgrade your oldest fab to the newest node, or build a new one, instead of shutting down a recent one and taking it out of production even though there is still demand for that node. There are still plenty of things being produced in fabs with older technology than that. Global Foundaries is the third largest in the world and they're offering 12nm or worse. People buy it because not everything needs a node which less than six months old and the price is right.
  - ksec 3y ago
    Yes. But that is specific to GF and TSMC when you have recurring customer on older node. Which is not true for Samsung Foundry.
    - AnthonyMouse 3y ago
      Why is it different? Wouldn't Samsung prefer to build new fabs rather than retrofitting old ones given the demand, since they would then have more customers and make more money?