Maybe, but AMD will market Zen as an enthusiasts "FX" platform, so unless AMD price it to compete with Intel's mainstream platform it will go up against whatever enthusiast SKU/HEDT Intel have. AMD could price down to Intel's mainstream, but that strategy hasn't worked particularly well up til now. Intel also have the option of dropping chipset prices back to X58 levels or lower (I'm sure the $50 they charge for X99 isn't warranted in any case) or offering a stratified chipset option if AMD look like achieving any serious inroads into Intel's business.
...
An interesting point. I always saw AMD as offering the better deal for upgrades, because your FX level processors and your standard processors all utilized the same chipsets (given, there was some distinction, but definitely not the fragmented mish-mash of Intel). I make this statement, looking at Intel right now. The PCH for SB and IB was basically the same offering (but requiring you purchase a new motherboard to insure compatibility with new CPUs and better PCI-e 3.0 support). Haswell finally got full SATA III support (and some M2), but that was about it. Broadwell never materialized. Skylake adds almost nothing to the mix that wasn't present in Haswell (at least nothing useful in the next 18 months based upon current trends and pricing). With each of these generations having at least 3 variants (and usually more), it was a pain in the butt to try and explain why one board cost more than another without having to whip out a comparative chart for features. Heck, that doesn't even cover the enthusiast lines and server lines (of which there was some cross-over).
While you're dealing with the Intel alphabet soup you had two or three chipset choices from AMD, that work with all their processors. If you bought a good motherboard when AM3+ came out it could see multiple processors over several years, assuming that you could live without some of the baked in features that Intel pushes with every minor revision. Right now I'm looking at several computer which have no real upgrade path (SB and IB), yet their PCH features are largely still sufficient for everything that I need today. Even if the CPUs weren't keeping up, the PCH would be enough for the next couple of years. That kind of logic works with AMD (or at least it did in the past), but Intel gives you 18-36 months (depending upon if there is a refresh).
I'm conjecturing that Intel only needs three cards to win most markets, and none of them is the raw performance card. They claim that they've got the most innovative features, by citing cost savings to businesses using their iGPU. They claim best thermal performance, to lock the server market. After that, the only card they need to play is recognition. Even if AMD went toe-to-toe, Zen couldn't beat these cards.
AMD has to win the enthusiast market for Zen to live, which wouldn't be hard with a sub $300 6 core or better CPU. Throw in all the PCH features that Skylake has, with an additional few PCI-e lanes, and you've got a platform worth spending some money on (it's easy once you admit that CPU upgrades will happen more often than new peripherals get added which will require drastic PCH changes). It's not like Intel gives a crap, based upon their enthusiast level PCH options (yes, I'm still stewing over the x79 PCH being both expensive and underwhelming). If AMD can make a decent dent on enthusiast CPUs they have a chance for their next CPU line to be a real success. I can't even count how many times over the past 5 years when somebody pointed out a cheap AMD system, and I immediately dismissed the savings because the performance wouldn't have justified it. If you can get people like us to seriously consider recommending AMD then they have a chance at being implemented. As it stands, I think the lackluster previous products prevent builders from recommending their product. They'll never get the server lines to survive, because you're right to say Intel will utterly stomp them with only minor changes to existing product. Cheaper CPUs are the only reason that AMD is still even in the CPU game, but that business is low margin and high volume (read: not enough money to support the R&D for future projects). I don't want to see AMD die, but even if Zen is a huge success it'll take a huge amount of effort to counteract the death spiral AMD management is making. Zen alone can't do that if they can't rebuild the loyal fan base.
I'm loathe to admit this, but AMD need people like SonyXperia. It needs die hard fans willing to buy day 1, and in order to do that AMD has to undo their reputation of mediocrity. The best way to do that isn't to compete with Intel, but to give those Intel has been ignoring a voice again. I can't be the only one angry that an ever increasing iGPU, lackluster generational performance increases, and consistently stupid choices (giving up solder, FIVR, etc..) has made buying a new Intel system feel like extortion over the past 5 years. Don't get me wrong, SB was the best overall platform I've ever seen. At the same time, IB, Haswell, and largely Skybridge have given me no reason to want to spend money on a platform. Assuming Zen isn't a flop for performance, I want to give AMD my business just to force Intel to get off their lazy backsides and make some real progress.
Let's also be people for just a moment here. Let's say Intel follows up Zen with Kaby Lake, and suddenly we see a generational improvement of 15%. I know that would piss me off to no end, because it'd be Intel telling customers that they only care about delivering their best products whenever there was competition. As a consumer, that's tantamount to being given the middle finger. Wouldn't that piss you off enough to take a minor loss in performance, just to give Intel the finger right back?
