If you've read the Jobs' biography, you know there's no way they're making that mistake again. The benefits of full control and cost cutting are small. Apple wouldn't want to alienate developers again. Intel does the r&d, drive the chip market and sells them at a relatively cheap unit cost. Apple has much bigger problems on their hands than their desktop/laptop processors.
Speaking of 'developers', I'd say by this point, more people are using the mac to develop ios apps and web apps than mac apps proper. All of those development tools and frameworks, still run on ARM. Also, of course, Java.
For mac developers, if they make use of mostly open-source libraries (and standard mac libraries) like most do, they'll be easily recompiled, as they have already, for linux, for instance. So, for typical apps, they can be recompiled for ARM.
----------
I have dozens of OS X apps and utilities I rely on every day. Even if Apple ported all *their* apps to ARM OS X tomorrow, the result would be a hobbled, limited ecosystem that would take years to recover.
It's not that it *can't* be done -- Apple has changed Mac CPUs twice already. But the benefit must clearly outweigh the cost and risk. What is the gigantic benefit with ARM OS X that justifies this cost and risk? If the transition took 5 years, during that period x86 is getting better in terms of performance and power consumption.
Besides the dual CPU idea of merging iOS and OS X, the only other reason I could see is if Apple foresees a *major* performance and power advantage with ARM. But many, many companies have gone down in flames betting against Intel.
Still it is theoretically possible. The x86 instruction set architecture and supporting microarchitecture are fiendishly complex, and getting more so all the time. Some researchers believe that only 10% of a modern x86 transistor and power budget is spent on execution. The other 90% is spent on supporting out-of-order superscalar methods needed to provide competitive performance. This is why a i7-4790K CPU burns 88 watts to produce 117 GFLOPS. By contrast a DSP chip using VLIW methods can produce higher performance at a fraction of the power.
It is conceivable Apple foresees Intel hitting a wall like Motorola did with the 68k family, and wants to start exploring alternatives. There are some researchers that believe significant exploitable gains still exist in single-core performance -- provided the instruction set architecture is changed. However many have felt that way in the past and look where Intel is today.
I'd imagine most of those apps are available on the mac app store? Apple has a lot more influence now with regards to mac app development, than they did before. Apple would pressure them to recompile. What made Intel great, is that they produced high-volume low-cost chips. which eventually allowed for their fabs to also become the best, yet if you look at the state of the world today, it's arm which produces the most high-volume, low-cost, power-efficient chips, and arm fabs are growing at a faster rate than intel, they still haven't caught up to intel yet, but the direction is clear.
----------
...or those developers just abandon their Intel Mac apps as many did with PPC-based apps, never bothering to update them to an ARM version. It has to be worth it to developers to go to the trouble. Within this thread, we're now envisioning some Macs with ARM and others with Intel. Maybe if Apple can sell huge volumes of these ARM-based Macs, developers might deem that worth the effort. How many Airs "as is" have been sold to date?
Or this goes back to the old question, why would people develop for apple computers anyway, when the market share is so small, and only develop for windows instead? Same for criticisms about the early ios app store, compared to windows mobile, etc. The reality is that it's vice versa. Companies build a product that people want, for whatever reason, in Apple's case, it's the experience and integrated ecosystem, and then people develop apps for the system.
