On Intel's (INTC 0.55%) most recent earnings call, company CEO Brian Krzanich admitted that the company was facing yield issues with its 10-nanometer manufacturing technology, pushing out the launch of the company's first 10-nanometer products (code-named Cannonlake) until the second half of 2017.
Although it's unfortunate that the company won't be able to deliver the performance and power benefits associated with the 10-nanometer process until 2017, there may be a slight benefit to Intel's desktop processor business as a result of this delay.
I know this sounds crazy, but I think you'll agree with me by the time you've finished reading this.
The mainstream desktop and high-end desktop platforms should be fairly well aligned
If you are a gamer looking to build a high-end, no-holds-barred gaming PC, then you are going to choose an Intel processor. However, even if money is no object, then you're faced with the difficult choice between the $350 Core i7 6700K and the $999 Core i7 5960X.
The 6700K features four of Intel's latest Skylake processor cores, running at 4 gigahertz base and up to 4.2 gigahertz at maximum turbo speed. The 5960X, on the other hand, features eight of Intel's Haswell CPU cores (two generations behind Skylake) running at 3 gigahertz base and up to 3.5 gigahertz max turbo.
With proper cooling, the 5960X can be overclocked to around 4.5 gigahertz and the 6700K should be able to hit 4.7 gigahertz (though the performance-per-clock of the 6700K is still higher than that of the 5960X thanks to the newer Skylake CPU cores).
Customers face a difficult trade-off here: Do they go for the absolute best four core performance at the expense of performance in workloads that can actually use eight cores? Or do they give up some on the per-core performance but go for double the cores?
To make the equation even more complex, the boards that the mainstream, lower core count CPUs often have better features thanks to the newer chipsets that those CPUs pair up with.
I suspect that there are folks who have the cash and the willingness to spend said cash on the $999 processor, but ultimately opt for the $350 chip because they aren't comfortable with the trade-offs that come with the high-end desktop chips.
Could this trade-off go away thanks to the 10-nanometer delay?
I believe that the gap between the launches of the mainstream desktop chips and their corresponding high-end desktop chips (which are derived from server/workstation processors) is set to shrink considerably as a result of the 10-nanometer delay.
Intel's Broadwell-E chip for high-end desktops is expected to launch in the first quarter of 2016, and the company's Skylake-E chips are expected to launch in the first half of 2017. Assuming that Intel stays on a cadence of yearly server/high-end desktop chip launches, we should see the first Cannonlake-E chips launch in the first half of 2018.
Given that Intel has said to expect the first mainstream Cannonlake parts to launch in the second half of 2017, this would imply just a six-month gap between the launch of the mainstream parts on a given architecture and the corresponding high-end desktop parts.
Right now, that gap currently stands at well over a year.
Why could this help Intel's desktop PC business?
The reason that I think that it's possible that this development could help Intel's desktop PC processor business is simple: Customers who prioritize per-core performance and platform features but are still interested in paying for more multithreaded performance if possible may now find the more expensive high-end desktop platforms more attractive.
That said, the extent to which phenomenon actually plays out (assuming, of course, the gap between mainstream and high-end desktop processors shrinks as I suspect it will) and what the financial impact will be, if any, is yet to be seen.
However, I think that if the launch schedules of Intel's mainstream and high-end chips are more closely aligned, the company's desktop business (in particular, its enthusiast/gaming processor business) should improve.
