> Fastest consumer Apple silicon versus fastest consumer Nvidia. So M5 Ultra vs RTX 5090. Not comparing price or efficiency or value or anything like that. Just the platforms as a whole. So M5 Ultra vs RTX 5090

The problem here is that they aren't equivalent platforms - no amount of money will buy you an RTX 5090 with 512 GB of RAM.

If you want to buy the 16x RTX 5090s you need to get to the same 512 GB of RAM, we're talking $100,000 just in GPUs, industrial levels of power/cooling/noise, and the interconnect bandwidth is still going to kill it for normal inferencing workloads.

Of course, that 16x cluster is an absolute beast in terms of raw compute - it's just not all that useful for the things people are trying to do with such machines today.