Resource constraints would be a concern, yeah, so if you were developing a game featuring LLMs (which would, at this point in their development and maturity, be a gimmick) you would keep that in mind and keep other demands on resources low.

True, yet many games are effectively single-threaded anyways.

The bigger problem is memory capacity and bandwidth, but I suspect folks will eventually figure out some sort of QoS setup to let the system crunch LLMs using otherwise unused/idle resources.