It seems that, for this to be true, AI (not just LLM) will have to plateau, very soon. Considering the Nobel prize for chemistry, last year, was about AI (not LLM) [1], I think there are still an incredible number of non-LLM applications of AI out there still.
[1] https://www.nobelprize.org/prizes/chemistry/2024/press-relea...
There are certainly periods where one concept is "viral" and appears quite often; that's normal.