Which doesn’t make sense, when the entry-level chips can outperform the supposed high-end chips. Given that reality, even less people will buy high-end machines, making them even less cost effective/more expensive.
What you say is not entirely accurate. The M4 chips do outperform M2-Ultra in peak performance on short tasks in single-core performance and neural engine throughput, they lag behind in multi-core and GPU performance. However, there is a wrinkle due to the thermal design of the different Mac models. Speed-based throttling in machines with an M4 chip, such as MacBooks or compact systems with limited cooling, significantly impacts performance relative to an M2 Ultra in a Mac Studio, which has a robust cooling system that enables sustained full-speed performance. In sustained tasks, the throttled performance of an M4 chip can be between 60% to 75% that of an M2 Ultra.
I saw the effects of speed-throttling clearly when I ran GPU-compute-bound AI tasks comparing my MacBook Pro M3-Max against my M2-Ultra. The M3-Max was about 80% the speed of the M2-Ultra. That's pretty impressive, nonetheless, but it means my M2-Ultra will still be more capable than other Mac models until it is replaced by an M4-Ultra.
I don’t see a reason why Apple can’t refresh the whole lineup within the span of a few months so that there is a clear value proposition for each product. The current rollout strategy is all messed up.
Logistics, engineering and manufacturing schedules, market demands, etc. I'm sure Apple sells a ton more machines with M and M Pro level chips than M Max and M Ultra chips. It makes sense for them to spend more time and effort on the much larger market for the affluent consumer than on the high-end content creator and software engineering niches.
Last edited: