Later this year, Intel (NASDAQ:INTC) is expected to release its eighth-generation Core processor family for personal computers. These chips, which should initially roll out for mainstream notebook computers as well as for enthusiast gaming desktops, are expected to be manufactured using the chipmaker's new 14nm++ manufacturing technology.
As the company disclosed back at its Technology and Manufacturing Day earlier this year, its 14nm++ delivers roughly a 10% performance increase over its 14nm+ technology (used to manufacture its seventh-generation Core processors) that Intel said offered a 12% performance boost over its initial 14nm technology, which was used to manufacture both its fifth- and sixth-generation Core processors.
Since Intel has made it plain that it aspires to launch new personal computer processors at an annual clip, it's going to need to follow up those eighth-generation Core chips with new products in roughly the third quarter of 2018.
Let's go over what those products are likely to be and what challenges Intel may face in bringing them to market.
The 10nm+ question
In the second half of this year, Intel is expected to go into production on its first-generation 10nm manufacturing technology, known simply as 10nm. Per Intel's disclosures at its Technology and Manufacturing Day, the 10nm technology will offer lower performance than its 14nm++ technology, but it promises to offer power consumption and chip area benefits.
Those products, which are based on the company's Cannon Lake processor architecture, are expected to begin shipping either by the end of 2017 or in early 2018.
Now, Intel does not plan to roll Cannon Lake out broadly across its personal computer product portfolio; it's going to be limited to thin-and-light notebooks. Desktop personal computers and higher-performance notebooks will be serviced by Intel's eighth-generation Core products.
What this means, then, is that in the second half of 2018, Intel is going to need to begin rolling out products based on its Ice Lake processor architecture for the desktop and high-performance notebook segments. It would also probably be in Intel's best interests to replace the Cannon Lake chips for thin-and-light notebooks with Ice Lake based ones in that time frame, too.
Intel hasn't disclosed much about Ice Lake, but there is evidence to suggest that it will represent a substantial architectural overhaul relative to the eighth-generation Core processors, as well as even the Cannon Lake chips (which Intel is apparently still trying to figure out how to brand).
I think it's likely that Ice Lake will be manufactured using technology that Intel dubs 10nm+, a performance-enhanced version of the original 10nm technology.
Here's what gives me pause, though: If Intel has struggled mightily to get its original 10nm technology into high-volume manufacturing (production start is nearly two years late from its original schedule telegraphed all the way back in 2013), the company's ability to deliver on a 10nm+ technology in time for product launches in the third quarter of 2018 is certainly questionable.
Right now, I'm watching for signs that Intel's execution in both chip manufacturing technology and product development will continue to improve. The company's recent pull-ins of Skylake-X for high-end desktops and Coffee Lake-S for mainstream desktops are certainly encouraging, but these products are manufactured on derivatives of the company's mature 14nm technology.
I want to see how Intel handles the transition to its first-generation 10nm technology (given the delays thus far, I'm not encouraged) and how quickly it can push out products based on performance-enhanced versions of that technology. If Intel shows obvious signs of improved execution over the next couple of product cycles, then my confidence in the company's manufacturing arm will improve.