However, I do think it's about time they wake up and start paying attention to Mac users needs because if weren't all pro users Apple wouldn't be standing here today.
Customer Loyalty = users dedicated to the manufacturer/brand
Manufacturer Loyalty to the User = Non existant with Apple (Not exactly customer service, but closer to customer support; it's rare in any industry though)
Seriously. If it gets in the way of the bottom line, it's gone. It's as simple as that, and the workstation market is a very small part of their total net profits per quarter. So it's possible it can happen.
The 2.8GHz quad iMac has a slightly faster processor than the base Mac Pro, it can't compete on graphics, IO or memory performance, nor does it have expandability. It can't compete with the processor upgrades available on the Mac Pro.
Overall, the MP has more to offer in overall performance, especially if the right upgrades are made.
But I see it as less to do with Intel (there are chips that can be used for workstations past 2013), but the market share Apple holds in this segment. We don't have solid numbers, but what is available, doesn't look good. Now combine Intel's price increases in this segment and Apple's margins, they're going to out-price their user base to the point it's no longer financially viable to continue using MP's. The sales numbers will drop to a critical point, and they'll have to pull it. That's not to say they won't try to use an iMac to replace it by leveraging LightPeak for graphics and storage, but it's lacking expansion which is a deal breaker for some at least. Resulting in more pros lost.
My GREAT hope with the move to Intel was that finally, Apple would have a reliable source of chips and thus would follow a SEMI-reasonable pricing adjustment schedule.
They do have a reliable source of CPU's, but the details are in other areas, such as low sales volume, normal development cycles, new products being introduced, and most importantly, a revolving development team. It seems they're cycled from project to project across the entire range of lines (no sub group dedicated to the desktop product range).
Completely nonsense for the current machines, but as nanofrog said, maybe one day the iMac will be as strong to be considered a workstation (even though it won't use workstation parts any more).
An AIO isn't ideal as a workstation, as many need the ability to add upgrades internally, and memory could be a real problem (some need more than what 2GB DIMM's can provide, as 4GB aren't the lowest or second lowest capacity yet). Even LightPeak could be made to bottleneck (even assuming it's real world performance = 10Gb/s <1.25GB/s>), a RAID can exceed that on it's own. Now consider improved graphics would have to use this as well, you see where it's going. That could be mitigated by using dual LightPeak channels, and assuming the final parts meet target cost, is quite doable ($10USD per in parts).
But given their recent products, it seems likely the AIO solution is what will happen from what it appears to me (history + minimal effort & costs).
Perhaps they'd make a new line (the real headless Mac that's been begged for for years), but only if it's able to deliver on desired levels of net profit in that segment. So low cost, not likely if it ever does show, as they'd want to make absolutely certain it won't cut into the iMac's sales, which is back to where we are right now, only with desktop parts (and that happens to be the direction Intel's going). But it all depends on the $$$, and it doesn't look to promising IMO.
TV). 
