A little while ago, respected technology website PC Watch revealed the existence of a future Intel (NASDAQ:INTC) processor family code-named Coffee Lake. This processor family, PC Watch reported, would essentially be a fourth generation of processors built on the company's 14-nanometer manufacturing technology aimed at certain segments of the personal-computer market.
Intel hasn't publicly commented about the reports, but a company executive may have accidentally revealed the existence of Coffee Lake at a recent investor conference.
Here's what was said
Intel's Brice Hill recently presented at an investor conference hosted by Pacific Crest. At the conference, an analyst asked Hill about the potential impact of the elongation of Intel's development cycle, from about two years to about three, on Intel's manufacturing lead over the rest of the industry, as well as the desirability of Intel's products to end customers.
Hill had quite a lot to say, but here's the bit that's relevant to Coffee Lake (emphasis mine):
"On product competitiveness, what you see with Intel currently is we're able to push the performance of our current generations of products. So for a short period of time, if we elongate the design cycle, we can develop a third or fourth generation of product that gives you a performance boost in the market that continues the customer expectation from a performance advancement perspective."
Intel may be employing a new strategy with respect to process
The main show-stopper with respect to Intel's 14-nanometer ramp-up was that it couldn't get manufacturing yields under control. Not only did this lead to an elevated cost structure for 14-nanometer products relative to their 22-nanometer predecessors, but it also led to some supply shortages, particularly for high-performance parts.
Intel's problem is that, historically, once it begins transitioning to a new manufacturing technology, it has to transition practically all of its products -- from high-performance desktop chips all the way to low-power 2-in-1 processors -- to this new manufacturing technology fairly quickly.
In the past, when Intel's process development went more smoothly, this didn't present much of an issue. Today, however, the reality is that Intel doesn't seem to be able to develop new manufacturing technologies and get them to high yield levels as quickly as it could previously.
Intel's new strategy could be this: Move to new manufacturing technologies as quickly as possible in segments that can really benefit from them, while building new products on more mature technologies for market segments that don't benefit as much from such technology transitions.
What such a strategy allows Intel to do is to keep a large portion of its product mix in more mature manufacturing technologies, minimizing the potential negative gross margin impact that would come from transitioning other products to a newer, lower-yielding manufacturing process.
Once yields on the company's newest manufacturing process matures, Intel can then begin releasing new products based on that process into segments of the market that hadn't yet undergone the process transition without fear of gross-margin erosion.