Issues with TSMC's 3 nm node may have forced Apple to adjust the A17 Bionic's performance targets. Initial rumours suggested the figure might be around 20%, but it is likely to be lower now. https://www.notebookcheck.net/Apple-A17-Bionic-s-performance-uplift-may-be-lower-than-initially-anticipated.702283.0.html
semianalysis did a review of TSMC's N3 node and it is definitely not progressing well at all. They described it as one of the weakest node updates in history. N3B is a joke and they said most companies would instead wait for N3E. I can't beleive Apple would be so stupid to use N3B given how small of an uplift it is in key metrics yet costs a lot more than the highly polished N4P node. Apple would be far better using N4P for iPhone 15 and waiting for iPhone 16 to use a much more refined N3P (assuming it's coming).
N3B (or original N3) and N3E have different design rules, so customers should redesign instead of just wait.
The N3B and A17/M3 are likely to experience similar hardships as Intel's Ice lake and Tiger lake. M3 Max may be delayed. Let's just pray that it doesn't turn into Whiskey lake.