By using this site, you agree to our Privacy Policy and our Terms of Use. Close
Cyran on 19 March 2023
Edit Thread
ConservagameR said:
Cyran said:

Here the problem, every single gamer could decide to stop buying Nvidia GPUs today and they still have a decent chance to have record profits over the next couple years.  With all the Hype around ChatGPT 4 and the many start up we starting to see with tons of money to spend that entire business model relies on licensing the use of ChatGPT and other AI models, the demand for NVidia expensive AI focus GPUs going to sky rocket.

All signs pointing to over the next couple years AI going to become a much bigger way of how businesses' operate.  Some of the recent Microsoft Copilot demos been impressive and it only going to get better and NVidia is in the best position to provide the GPU to run these AI models.

Just look at articles like

https://www.digitaltrends.com/computing/microsoft-explains-thousands-nvidia-gpus-built-chatgpt/

This looking to be even more profitable and more sustainable for NVidia then the mining boom.  

Then why hasn't Nvidia bailed on gamers already if it's such a waste of their time?

Misinterpret my point.  I don't believe Nvidia going to ever leave the gaming market or stop putting RD&T.  What I am saying is that they will focus on features that can be used in more none gaming GPU also and if that mean they price out a lot of people they not that concerned.  They more interested in selling the most powerful GPU to enthusiast then designing a GPU for the masses.

This always been the case but starting with Kepler they was very focus on the mobile market as you see with all the shield devices they launched which meant a focus on energy efficiently and the GPU architectures they released show that.  This happen to work out well for gamers and pricing.

By the time the 20 series release there focus shifted too AI and therefore there gaming GPU also reflected that.  I actually agree ray-traying was more of a well we spent all this RD&T on AI focus stuff how can we also make use of it for gamers.  The fact that packing all this AI stuff into the GPU was going to increase cost of there GPU was not something they concerned with. 

You can also see this with the 30/40 series in that they know the most units they sell is usually the xx60 lines and in previous generation they release that 2-4 months within the flagship.  In 30 series it was a entire year and 40 series we still waiting on.

I just don't believe Nvidia cares if people not willing to pay $600+ for a GPU stop buying there discrete GPU if the architecture requirements for there AI GPU is expensive that going to reflect in there gaming GPU prices.

They not going to make 2 completely different architectures so the data center and gaming one going to share a lot of stuff and the focus always going to be on what make the most money.