They probably can go to a yearly release (they do it with the iPhone, and they're related designs) - but they can't do BOTH yearly releases AND "every release is a big deal".
They may end up doing something like "chips release yearly, but desktops only get updated every other chip". The danger there is something being off-cycle. For example, a cycle where the base chip is the main beneficiary, but the iMac is already recent and the Mac Studio needs an update.
Or, as just happened, a Max-centric cycle, but the Studio and Pro were just released (they REALLY should have held those machines 4.5 months, skipping the M2 generation). There may be an embarrassingly long gap when the M2 Ultra models are "current", but no faster than the M3 Max MBP on many tasks. Even if they get those machines out at WWDC, about as early as they could, they will have existed as "ghosts" (current models, but clearly superseded) for twice as long as they were really current. The M2 wasn't even a big update over M1 for the big chips - the Studio and Pro could have easily waited for the right chip.
They'd need to get very good at coordinating their chip design efforts to avoid this (and they aren't right now). If they really had control of cadences, they could do something like this:
Year One: Chip that is based on making the individual cores faster and on efficiency, with no major changes in core counts (maybe some extra e-cores somewhere). Lead with the MacBook Air and either the iMac or the Mini right after the iPhone. MacBook Pro and the other small desktop in January by press release. No update to Studio or Mac Pro.
Year Two: Chip that is focused on the "big" versions, adds P-cores or significant numbers of GPU cores to the Max (etc). Lead with the MacBook Pro right after the iPhone. Big desktops either join the MBP or aim for January. Air squeezed into an iPad launch. No update to iMac or Mini.
That leaves them with laptops every year (although on an odd cadence with some short-lived models), desktops every two, and everything getting the most relevant chips.