The root cause is missing the mobile phone revolution. Everyone’s favorite, most used computer is in your pocket. That used to be the laptop PC. Phones replaced giant swathes of PC functionality (email, photos, etc) and added new compute+mobility features not possible on PCs like Uber, instant photo sharing and so on in a more convenient pocketable form factor.
It expanded the market from 200-300M PC annual units to multiple billions of mobile units. All of those phones demanding ever more compute on leading edge nodes fueling the economy of scale of TSMC.
Intel saw the threat too late and made a bunch of bonehead moves akin to a chicken running around with its head cut off. The truth is that Intel could never embrace mobile early enough due to misaligned incentives (replacing high margin PC silicon with low margin mobile silicon).