...When it debuted, it was relatively competitive to similarly equipped workstation PCs.
debuted as in the 2006 model. Yeah that is probably an artificial low as that was largely just taking an Intel motherboard and pushing it out to market as fast as possible. ( the objective was to flip the whole mac line up to Intel in a year and Apple doesn't have that much design bandwidth. )
The core problem they have though is that alot of customers price anchored on that. Perhaps a small part of the major product change is to help reset the anchor but
Cost could be saved by not having to include support for the workstation class chipset, no ECC memory, only 1 thunderbolt controller, no second GPU and the entry level CPU could be something like a Core i5 around $200.
A > $2,000 for a i5 system? I don't think Apple "value add" is enough to cover that gap except for the most deeply religiously attached. Apple could gut the internals to maximum extent but at some point folks are looking for value at $2,000+ and if it isn't the hardware in the system .... what is it they are looking at?
Apple tried that back in the PowerMac days. They would sell some "gutted to the max could reasonably get away" version just to hit a $1,500-1,700 price point. That was more so Apple selling a "bare bones" system (at least as bare bones as Apple would go.). Many folks would partially empty it and use it as a baseline for they new "erector set" project. Those tactics and strategy didn't build a 10+M/yr mac market. It is pretty clear Apple has moved past the "come throw whatever you want inside of our system" stage. It doesn't really work for what they do best.
A $2,000 'bare bones" box probably isn't going to work. Even if put in a slot for the box-with-slot lovers ( they typically have their own price anchors which are based on the overall class PC component market). A $2,000 "bare bone" system that is heavily aligned with the current Mac Pro would work even less effectively. Basically copying the Mac Cube at that point. Under horsepower (relative to other "desktop" class Macs in the line up) and crippled to hit lower pricing and largely closed system. There is already a track record for that.
If there was a new desktop Mac coming it seems more likely that it would be a a Mac micro split off of the Mac Mini than something to fill the gap between the iMac and MP.
This was true for Haswell. For Broadwell, consumer chips were released Q3 2014, and Broadwell-EP will be released Q1 2016.
Broadwell consumer chips didn't make 2014 at all. That was the logjam problem. They were suppose to but didn't. Once Intel finally worked out the 14nm process problems the desktop Gen 5 Broadwell products were getting close to the Gen 6 ( Skylake) ones. That's was the root cause issue. That's why Intel has the "Haswell Refresh" in 2014 to attempt to mitigate some of the damage ( move the chipset up to Broadwell compatible ).
For Xeon E5 there was no and never been hinted at any "Haswell Refresh". Intel could play some hocus-pocus name game. Just change v4 ( "Broadwell-EP") into 'v3.5' ( "Haswell-Refresh-EP") . We'll see. They are socket and chipset compatible. [ 'v3.5' would be v3 with some new intermingled model numbers. ]
If you want Skylake-EP, you are going to be waiting 12-18 months from now.
I can see Intel moving v5 out of 2017 and into Q4 2016 (and pre-announcing and leaking heavily in Q3 ) . Especially if that means stomping on AMD's Zen offerings announcements that are oriented toward servers.
