I'm not saying that Apple has a monopoly in anything, but the sheer amount of volume they churn out each year in iPods, iPads, and iPhones is enough of an excuse for them to go into the fab business. The rumors have persisted they'll eventually move their macbooks to ARM processors once performance is good enough (which makes sense as it would give their entire product line backwards compatibility from a hardware standpoint), which would further increase the volume.
Makes me wonder...Does apple sell more iPads, iPods, Macs, and iPhones combined than Intel sells CPU's nowadays?
The economics of owning your own fab requires you to have a business that entails the production and continued sales of legacy products.
If Apple were to buy a fab that was equipped with all the required tools for 22nm production, that fab needs to still be producing 22nm chips 6-8 yrs later in order for the financials to make sense.
This is why a foundry owns the fab but the fabless customer does not. Apple is not about selling legacy products, they are about getting people to upgrade and feeling like their perfectly working 2nd generation iPad needs to be replaced with a spanking new 3rd generation iPad and so on.
A typical fab will add on new nodes to its production load as time goes on, but it rarely phases out the production of legacy nodes. For a foundry this is not an issue because there are always a supply of small fabless companies out there who trail the leading edge by 4 or 5 nodes for cost and complexity reasons.
But what would apple do with a 22nm production line in 6 yrs when they want to be have all their latest gizmos and gadgets using 7nm? If the conclusion is going to be "just shut down all 22nm production and requalify all the tools for 7nm production" then they may as well remain a fabless company that throws gobs of money at a foundry to ensure the foundry has the ramp-to-capacity that Apple needs.
(the economics there are in Apple's favor, they are not fab experts, they are gadget and software experts, the foundries are the fab experts)
To own a fab pretty much requires you to have a desire to have a product portfolio that entails producing legacy and lagging-edge products for years and years. This works for IDMs because they force themselves to adopt a business model that enables it - Intel and AMD getting into the chipset business for example, or foundries being able to support lagging-edge fabless companies, or businesses that do catalogue sales such as automotive or medical parts that need be produced for 10yrs or more.
Apple does not fit that mold currently.
Now maybe they have grand plans for a shift in their business strategy itself, and with that might come the opportunity to sell large volumes of legacy products for years and years down the road such that owning a fab begins to make economic sense?