On the tech side, I think work will split on two tracks: 1) building great applications with small and medium fine tuned models like Mistral, etc. Within a year or two great models will run on the edge because of continuous technical improvements. 2) some players will go for the long game of real AGI and maybe they will get there in much less than a decade.
On the business side, I have no idea how the current situation is going to shake out.