Originally Posted by Sazar
You have to develop different drivers, software apps and other items to run specialized cards. There is a related cost for R&D. No doubt the margins are much richer for these cards as well, but there is a reason that the added costs are there is because it is more expensive to produce.
Well yeah obviously there are other costs on that side, but thatīs not what the quote I was commenting was saying either...
I was referring to the part that its "old news" that workstation and engineering/scientific applications costs more, I wasnt presenting that as "news" to anyone, as the quote seemed to imply, that wasnt the point.
But yes, the costs of investments is a good point.
For example, the entire effort sunk into CUDA and GPGPU is very much an investment to build a market for Tesla, get people using the API and GPGPU applications in general, on as many fronts as possible, get it tought in universites etc, they touched on that on GTC aswell.
Spending to get a return on the boards, I get that.
So then the margins on that stuff is already tied up in recouping investments perhaps.