To sum it up, you may also ask why Xeon instead of i7, and the same logic will apply.
Added price accounts for stability, lesser power consumption and improved drivers for professional applications. That doesn’t mean that you couldn’t get a consumer grade (GTX) card to achieve the same if not better performance, but your electricity bill, noise levels, and stability would suffer greatly.
The professional alternatives aren’t meant to be “faster”. They are actually often times clocked at lesser speeds exactly for that very reason: lower voltages, less heat, more stability.
Also, S_D’s argument is also valid, but only if you are actually a professional 3D artist and need to load that sort of huge texture information into your VRAM (large level designs, at once).
There’s no magic answer there… you want professional reliability, you buy a professional, more expensive card. You want the best performance for your buck, disregarding the odd crash, go for the consumer one.
p.s: forgot to mention, unless you’re talking about GFX based rendering (octane/arion whatever else works) you still depend entirely on your processor, and the v-card will only be used for GUI and viewport animation, so the cuda core count outside of that usage scenario is about as relevant as the color used on the PCB