Once on Skylake, Intel could run it for 3-4 years while turning the full focus and attention of the company to solving the mobile problem and developing a massive position in solid state drives.
Fischer also points out in the article that Advanced Micro Devices (NASDAQ:AMD) is "basically history in the processor business," and suggests that Intel can milk the Skylake design in PCs while devoting resources to other businesses (such as solid state drives and mobile).
I respectfully disagree with Mr. Fischer.
3-4 years is a long time
In the PC chip market, 3-4 years is a very long time. If Intel were to launch Skylake in the third quarter of 2015 and let it run as its main PC product until late 2018 to late 2019, then this would give AMD more than enough time to "catch up" both architecturally and on manufacturing technology (foundries should be well into 10-nanometer by late 2018 and certainly at 7-nanometer by late 2019).
Additionally, remember that article that I published earlier this month describing how AMD managed to get some architectural punches in on Intel with the former's Carrizo vis-a-vis the latter's Broadwell chip? Even though Intel's chip is on a far more efficient manufacturing technology, AMD included specialized HEVC video decoding blocks (which improve efficiency) that Intel's Broadwell doesn't have. Intel brings this with the Skylake chip in the second half of 2015.
Now, imagine this: Intel "hangs back" to milk Skylake, but in that time, AMD continues to advance its architecture and moves to new manufacturing technologies. Even though Intel is far better capitalized and has many significant advantages, I'd bet money that AMD would have the ability to catch up if Intel were to be the proverbial hare to AMD's tortoise.
It's not just AMD, either
In case you haven't noticed, the semiconductor manufacturing technology at the major foundries is being driven by mobile technology. Think smartphones and tablets. So, what happens if Intel keeps Skylake on 14-nanometer, but -- as Fischer suggests -- focuses all of its energies on mobile?
This would imply that, by the late 2018 to late 2019 timeframe, Intel will have ticked and tocked its way to 10-nanometer or even 7-nanometer Atom products. These Atoms would likely offer better performance than Intel's Skylake in thermally constrained form factors. Indeed, Intel's smartphone chips at that point could very well offer more grunt than their best 15-watt laptop chips. Same goes for Intel's mobile competition.
Oh, and I'm sure if Intel decided to stagnate on Core, Apple (NASDAQ:AAPL) would show Intel the door with respect to the Mac products.
So, no, this isn't happening
I fully expect Intel to continue launching products as quickly as technically (and economically) feasible. Not only does Intel benefit from new manufacturing technologies, but new architectural features. Think better graphics, support for more robust video decode/encode, faster processors, and so on. Intel needs to keep pushing on new features to keep the PC interesting; slowing down its chip development/release cadence would, in my view, be disastrous.