@mhhfive
Perhaps, but I don't really see Tim Cook as the man to push Apple to boldly enter new markets.  He is more of a quarterly upgrade sort, he has yet to exhibit the kind of vision Steve Jobs did, the kind that creates brand new products.

Take the two big "new products" of his tenure, thus far:

iPad Mini
Retina MBP

Both are really not "new" at all, they're just riffs on the preexisting product, with some small tweak.

If Apple is going to one day conquer the server market, it will probably need a new chief.

Mitch: I dont think Apple has slowed down as such. True there is some delay in big developments after the demise of SJ but surely they have the talent to back their business for sure.

Since Steve Jobs's return, Apple has shown little interest in any but the most casual integration with the enterprise market. They'll add some features for the enterprise -- management tools, Exchange integration -- but nothing that requires heavy lifting. It's just not in their business model. 

While we both agree that it fits with Apple's culture that they're probably trying to use their own custom hardware for their datacenters... it wouldn't shock me to find out that they're using Oracle hardware/software for the stuff that they really have no business developing themselves. I doubt Apple has developed its own enterprise-level database system to handle all of its itunes transactions... but maybe it has? who knows.

But if Apple *has* developed its own enterprise datacenter hardware/software... there might be a decent market for it somewhere? Pixar maybe? 

@Mhhfive

As for Apple, I think Apple's approach is largely driven by margins.

Server hardware and software is a money maker and has better margins than some sectors, but still pales to Apple's gaudy margins on its luxury computers/mobile devicees.  While prices on said products have come down somewhat, Apple has brilliantly (or abusively, depending on your perspective) squeezed both ends of the supply chain (manufacturers, carriers) to the extent that it makes way more money than anyone else in the industry.

That approach may work with consumers who are heavily swayed by looks and emotion, but I think Apple realizes that it can't razzle dazzle enterprise users into paying some exorbitant markup for the same product, as they tend to be savvy buyers.

Thus I think the approach that makes the MOST sense for Apple financially is to build its own optimized server hardware in-house as a means of supporting its high margin products; that's likely what it is doing already, as you mention.  There's not much to really gain from devoting time, R&D money, and energy to pushing out lower margin product to compete with Dell and HP in the enterprise server market.

@Mhhfive
Yes in theory x86 could be licensed, the question is who would do it.  Intel -- for now -- clearly does not want to travel down that road.  Maybe someday it will, but for the time being it has no announcements towards that end.

Also, it would be inherently more difficult to modify x86 cores as they have a larger instruction set and are hence slightly more complex to optimize.

This difference is eroding slightly, as ARM Holdings tacks on more capabilities to its cores, but it still is substantially easier to tweak an ARM core than an x86 one.  From my perspective, that's a key reason why AMD is making the switch to ARM (and yes, I do see it as a switch, versus a diversification, most likely) -- it just isn't cost effective to design such monolithic cores unless you have a massive R&D budget to do it right (like Intel).

For Intel, x86/in-house design makes sense.  It will continue to produce great, optimized cores, albeit with less flexibility.

For Samsung, AMD, Qualcomm, NVIDIA, and just about everyone else, licensed RISC cores make more sense from a price v. performance standpoint than trying to somehow jumpstart an effort to license monolithic CISC cores or to design their own CISC/RISC cores.

I expect Intel to continue to do well in the server market for years to come, but there will be cannibalization as the ARM proponents (pretty much every OTHER major chipmaker in the industry) leverage their advantages to gain ground in this new arena for the RISC architecture.

One last thing to note -- keep your eye on MIPS.  Imagination Technologies' decision to by MIPS chipmaking unit and produce licensed IP cores based on this veteran architecture could make it yet another viable RISC competitor to Intel.  Essentially the same advantages that apply to ARM co-designed cores should, in theory, extend to co-designed MIPS cores.

Just look at the target market for OS X Server... Apple's own website describes the Mac Mini as perfect for small businesses and hobbyists. 

I'm sure someone out there might be able to hack a mac mini into a super-performing BSD server, but OSX Server isn't really meant for huge projects. So as the way Apple is currently selling it, it's not aimed at large enterprises. 

Maybe someday Apple will reveal a "one more thing" and tell the world that it's been running iCloud on a datacenter filled with Mac Minis running a previously uncommercialized OSX Server:Enterprise edition... but I'll believe it when I see it. 

Apparently, you can install Oracle on OS X, but maybe not the latest version of Oracle.... so I think that says something about the acceptibility of Apple servers in an enterprise setting.

mhhfive - Interesting. What are its shortcomings as an enterprise server?

Given that ARM servers are a nascent market right now, I'd guess that any company that's experimenting with ARM-based datacenters is doing it for the "right reasons" -- and the increased management overhead is worth the potential gains in performance and energy savings.

You don't see the Mac Mini is a competitive server?

Uh, no. I realize that people use Mac Minis in colocation facilities and whatnot, but it's not a serious competitor to other enterprise hardware in major datacenters. I seriously doubt Apple itself uses Mac Minis for its own datacenters...