Image source: Intel.
In a previous column, I wrote about an interesting job listing that I found on Intel's (INTC +2.86%) website. In particular, it seemed to suggest that products based on the company's upcoming 7-nanometer manufacturing technology might not arrive until either 2021 or 2022.
I recently stumbled across something, hidden in plain view, that seems to corroborate this idea. Let's take a closer look.
Check out this slide from Intel's developer forum
At Intel's recent developer forum, two executives gave a presentation disclosing some key details about recently announced and future chip manufacturing technologies, as well as the company's strategy to provide contract chip manufacturing services to third parties.
In one slide, Intel showed how it intended to offer three different iterations of its 10-nanometer technology known as 10-nanometer, 10-nanometer+, and 10-nanometer++. Each successive iteration is expected to deliver better performance "to support multiple leading-edge products."
Looking more closely at the slide reveals something interesting:
Image source: Intel.
Notice that the circle denoting the third iteration of the company's 10-nanometer manufacturing technology, known as 10-nanometer++, actually sits solidly beyond the 2020 time-frame.
This could mean one of two things
If Intel really isn't planning to get its 10-nanometer++ technology out there until sometime past 2020, then this could mean one of two things.
The first, and certainly the more pessimistic, implication is that the company's 7-nanometer technology won't arrive until well beyond 2020. Indeed, if 10-nanometer++ is expected to arrive in 2021 or even 2022, we could be looking at late 2022 or early 2023 before the chipmaker transitions to its 7-nanometer technology.
Such a significant delay, particularly in light of the fact that contract chip manufacturer Taiwan Semiconductor Manufacturing Company (TSM 1.55%) plans to put its 5-nanometer technology into production in 2020, would mean that Intel will fall behind its biggest competitor in terms of chip area scaling. (TSMC's 5-nanomater technology is likely to be similar in terms of transistor density to Intel's 7-nanometer technology.)
An alternative explanation is that Intel actually intends 10-nanometer++ and 7-nanometer to co-exist. For applications that absolutely need to be on the latest, densest manufacturing technology, Intel could design products on the newer, and potentially more expensive, 7-nanometer technology. For products that don't need to adopt bleeding-edge manufacturing technologies -- for example, desktop processors -- 10-nanometer++ could be a better, more cost-effective choice.
In fact, we are already going to see something similar to this next year. Intel's next-generation desktop processors, code-named Coffee Lake, are expected to be manufactured on the company's 14-nanometer technology, while the low-power notebook and 2-in-1 processors, code-named Cannon Lake, that are expected to arrive at the same time will be built on the company's newer 10-nanometer technology.
Some additional clarity needed
At Intel's investor meeting later this year, I expect the company to go into more depth on its chip manufacturing strategy and how it views the competitive landscape. It will be interesting to see if the company talks at all about the potential timing of its 7-nanometer and potentially even its future 5-nanometer technologies.
I'm not going to hold my breath for such disclosures, although given how aggressively the competition is advertising its technology plans, they would be most welcome.
