I have 2 Gigabyte models over ConnectX-7.
recently started getting into ai, around 2 months back. got hooked on the idea of having everything local so started off with 16gig gpu, upgraded to 5090, then got single dgx spark and then got asus ascent gx10 as well. so currently running deepseek v4 flash on 2x gb10 nodes and qwen3.6-35b on the 5090. developing apps for my business
I have a spark and an ascent, running as a 2 node cluster.