But that's not how it would work. For one thing, launch units must ship by a month before the date, to make it to shelves on time. So the 2.74m were made in 3 months, not 4. However, this was during months when Apple was not hogging components. If Nintendo started in March 2016 for November they'd have more time, but they'd also be producing less per month, just like they are now in spring and summer 2017. Instead of 3 months at ~30k per day, they'd have 6 months at ~22k per day. That would increase launch to a putative 3.96m, and provide an extra 750k in the following quarter (with no Apple competition). But then the imaginary third quarter would be lower than real Switch, somewhat mitigating the advantage.
All in all, they'd get about 1.7m more. But as I said, this would be pretty meaningless overall. They'd still be behind PS4 (and now with no "but it didn't launch at holiday" rationalization available), and within a year they'd be back behind Wii, GBA, and DS again.
...In theory. This is all very tenuous projection, and I wouldn't put much stock in it. But I also don't think there's a more plausible version where they could achieve the "extra 2m at launch, extra 3m overall" level you suggest.