As we head into this week's CES, we're expected to learn of Intel's brand-new Kaby Lake desktop processors. That leads us to ask a big question: what's next for Intel? Kaby Lake, while undoubtedly Intel's most advanced architecture to date, is a bit of an oddball in the company's portfolio as it's neither an official "tock" (architecture update) or "tick" (die shrink), due to the fact that it's been really tough to shrink these chips down further. That's the crux of all the doubt that surrounds the future of Moore's Law.
For the time-being, though, Intel hopes to proove that Moore's Law isn't dead, as it's expected to release its first 10nm chips in 2017. To put give some perspective on this timeline, Intel's first 22nm desktop processor came as Ivy Bridge in the spring of 2012, after which point it took 2 years to greet the launch of 14nm Broadwell. When Cannonlake arrives, sometime later this year, it will have taken 3 years to see another die shrink, to 10nm.
Cannonlake, like the initial Kaby Lake mobile chips, are going to target the low-power segment of the market, due to weak 10nm wield at this point in time. That being the case, the direct follow-up to the upcoming Kaby Lake desktop chips will be Coffee Lake, expected to launch in the first-half of 2017. Unlike Cannonlake, Coffee Lake chips are going to be an evolutionary follow-up to Kaby Lake, sticking to the 14nm process.
According to industry reports, while 10nm wafers are going to be more expensive to produce over 14nm ones, the resulting cost per transistor will be lower. And if you think for a moment that Moore's Law is in fact slowing down, you can know that Intel Senior Fellow Mark Bohr disagrees, stating, "One important message is that this node, and the products that we’ll be making on it, will hopefully dispute some of the concerns of the industry that Moore’s Law is slowing down."
For us tech fiends, that sounds great to us.