In mid 2015, website BenchLife.info leaked to the world that Intel's (INTC -0.12%) 2016 processor family would be a series of chips based on the company's Kaby Lake architecture. Prior to that leak, the technology and investment community had expected that Chipzilla would introduce its first 10-nanometer processor architecture known as Cannonlake in 2016.
The big surprise surrounding Kaby Lake is that it is built around the same 14-nanometer technology that Intel's previous two architectures -- Skylake and Broadwell -- were, signaling a slip relative to the company's traditional two-years-per-manufacturing-technology cycle.
Management has told investors that they are pushing to try to get back to a two-year cadence post-10-nanometer (presumably they mean a two-year transition from 10-nanometer to 7-nanometer), however, from what I have just learned from a source familiar with Intel's plans, the company is working on three, not two, architectures for the 10-nanometer node.
Cannonlake, Icelake, and...?
Intel has officially said that its first 10-nanometer architecture is known as Cannonlake. The second 10-nanometer architecture, which is expected to launch about a year after Cannonlake, is known as Icelake. The follow-on part, I am told, is known as Tigerlake, and it will be built in the same 10-nanometer manufacturing technology that Cannonlake and Icelake are.
What does this mean for Intel's manufacturing technology road-map?
If we assume that Intel launches Cannonlake in the second half of 2017 as the company has said, Icelake in the second half of 2018 (one year after Cannonlake), and then Tigerlake in the second half of 2019, then this would mean that the company won't launch its first 7-nanometer processor architecture until the second half of 2020.
If all goes well for the company, then 7-nanometer could be a two-product node, implying a transition to the 5-nanometer technology node by the second half of 2022. However, the source that I spoke to expressed significant doubts that Intel will be able to return to a two-years-per-technology cycle.
How does this compare to what manufacturing rival Taiwan Semi is claiming?
In my view, Intel's fiercest competitor in the world of chip manufacturing technology is contract chipmaker Taiwan Semiconductor Manufacturing (TSM 1.60%). Taiwan Semi claims that it will go into high-volume manufacturing on its 10-nanometer technology (which should have slightly better density characteristics than Intel's 14-nanometer technology) in late 2016/early 2017.
Taiwan Semi then plans to begin mass production of its 7-nanometer node during the first half of 2018 and expects to move to its 5-nanometer node "about two years after [its] 7-nanometer."
At this point it is starting to look as though Intel could very well lose its manufacturing leadership relative to Taiwan Semi by the time TSMC transitions to the 7-nanometer node, and then TSMC could even extend this leadership at the 5-nanometer node (depending of course on how the electrical and physical characteristics of TSMC's 5-nanometer and Intel's 7-nanometer nodes compare).
It's time for Intel to start being more transparent with investors
As an Intel shareholder, I am deeply saddened that management is seemingly becoming less transparent with investors when it comes to chip manufacturing technology. From where I'm standing, it would appear that management makes statements that are not untrue, but can be interpreted in such a way so as to put Intel in a potentially "too favorable" light.
For example, management showed the following slide at its most recent investor meeting:
Image source: Intel.
At first glance, this slide appears to tell investors that Intel has a substantial lead over the competition. There is nothing untrue on this slide -- Intel's 14-nanometer node is substantially denser than Taiwan Semi's 16-nanometer node, and I have little doubt that Intel's 10-nanometer node will be substantially denser than TSMC's 10-nanometer node.
The problem is that this slide is not particularly useful without additional context. For example, Taiwan Semi is planning to go into mass production on 10-nanometer in the fourth quarter of 2016 and is planning to start mass production of 7-nanometer in the first half of 2018.
If we assume that Intel goes into mass production on its 10-nanometer in the first quarter of 2017, then this would imply that Intel falls behind for a quarter before "leapfrogging" Taiwan Semi. However, TSMC has an opportunity to catch up to, if not actually outpace, Intel once the former begins 7-nanometer manufacturing.
Given that it is not likely that Intel will transition to its 7-nanometer technology until sometime in 2020, Taiwan Semi will again have the opportunity to catch up/outrun Intel with its own 5-nanometer node, which looks as though it will into mass production around that same time as Intel's 7-nanometer technology.
If Taiwan Semi actually delivers upon the timelines that it has signaled to investors, and if Intel stays on 10 nanometers for three processor generations, then it's not at all obvious that Intel has any sort of manufacturing lead over TSMC, which manufactures chips for the vast majority of Intel's direct competitors.
