I would not put it beyond Apple to have the MacBook Air regularly trail the MacBook Pro by one SoC generation for up to a full year, possibly even longer, from the moment the ("base") MacBook Pro gets upgraded to the M6.
For one, you can see a precedent for this strategy in iPads, in when the iPad Air is updated vs. the iPad Pro. Also, for the MacBook Air specifically, a similar delay is already being practiced, just not to the same extent. Since the M3 generation, the MacBook Air usually trails the Pro by about 4-5 months already.
Whether Apple increases this delay will be a matter of competitive pressure or rather lack thereof. The less competitive Intel and AMD are in the same general price category, the more likely Apple is going to get away with it.
On the MacBook Air, this would keep the BOM lower (and margins up), especially when the latest SoC generation moves to a new process node. In that situation, it also enables additional purchasing flexibility for capacity on latest-gen nodes, by delaying some of the volume until a year later, enabling a more gradual ramp-up, more stable yields, and more viability in slipping in a smaller early order (and, more importantly, a smaller later order) on a latest-gen node when competing for capacity with AI hyperscalers. It also helps to further differentiate the MacBook Pro from the Air, which Apple usually loves doing, thereby selling a larger quantity for a premium. For M-series SoCs, differentiating by SoC generation is more economical compared to developing an additional base-chip SoC variant for the Pro, like, e.g., the A19 Pro vs the base A19. For the A19 Pro, a separate die design is more economical because iPhones significantly outsell devices with M-series SoCs in quantity. Being smaller, the A-series SoCs will also have higher yields, restricting binning opportunities.
