Why Intel Corporation's 14-Nanometer Mobile Chips Could Finally Be Game Changers

It has been a frustrating several years waiting for Intel's (INTC -1.79%) mobile strategy to start paying off for shareholders. On the plus side, the progress Intel's cellular modem group is making is phenomenal, with the gap between Intel and Qualcomm's (QCOM -1.75%) comparable modem technology measurable in quarters rather than years. On the negative side, Intel's mobile applications processors have been, to borrow a phrase from Sanford Bernstein's Stacy Rasgon, "a horror show."

Going into Intel's investor meeting, I was looking for a compelling mobile system-on-chip roadmap for 2016 and beyond, and for the first time in a long time, I think Intel finally has one in place.

Here's how the market leader does it
One of the reasons Qualcomm has been so successful is that it has what it calls a "multi-tier roadmap." This might sound fancy, but all it means is that Qualcomm designs system-on-chip products that are price/performance/feature appropriate for their target market segments, as illustrated in the table below:

Source: Qualcomm, author analysis.

Intel did not previously have such a roadmap. What it did is try to develop products for the high end, but since those products were late, it then tried to pit them against the mid-range and even value offerings from its competitors. Unfortunately, this proved to be an ineffective strategy, as Intel's products didn't integrate everything they needed to, and often had serious platform bill of materials cost issues.

In other words, yesterday's high-end processor isn't necessarily today's optimal mid-range or low-end processor. This finally changes, though, as Intel has signaled that it will have a top-to-bottom, properly targeted product stack in 2016.

Two SoFIAs and a Broxton make for a complete product stack
For 2016, Intel has signaled that its mobile product stack will be as follows:

Source: Intel.

While Intel didn't give much in the way of technical details for these chips, the numerous leaks that have hit the Web suggest that Broxton will be a stand-alone applications processor built on Intel's 14-nanometer manufacturing process and feature Intel's Goldmont processor cores. The SoFIA MID and SoFIA LTE 2 will likely be integrated products also built on the company's 14-nanometer process, but further technical details remain unknown.

While I do think building the entire stack on 14-nanometers will afford Intel a manufacturing lead, and perhaps performance lead, over competitors, the important thing here is that Intel has a complete stack that it can offer to its customers. This means Intel can be a one-stop shop for smartphone and tablet OEMs looking to build out a family of products, as Qualcomm is today.

But, competitors will be on FinFETs, too
It is possible -- some would even say likely -- that by the time Intel's SoFIA LTE 2, SoFIA MID, and Broxton are fully ramped, Intel's competitors may have moved to 14/16-nanometer manufacturing technologies. I have my doubts that the mobile processor vendors will move to 14/16-nanometer technologies before mid-to-late 2016, but let's assume the competitors do get there -- even for their bottom-of-the-barrel budget chips -- for the sake of this analysis.

Quite frankly, it probably wouldn't matter.

If Intel designs its chips and platforms correctly, then it should be able to offer very competitive performance and a very attractive cost structure. 14-nanometer Intel parts launched in 2016 will benefit from the fact that Intel has been in mass production on 14-nanometer since the second quarter of 2014. As a result, yields should be fantastic, and I would be shocked if Intel didn't have the lowest cost structure in the industry for a 14/16-nanometer technology.

If Intel can deliver best-in-class platforms in 2016, then I think it will have a cost structure advantage that will allow it to profitably accelerate share gains in the smartphone and tablet markets.

Intel needs to execute, though
Keeping this thesis together is the assumption that Intel will actually execute. This has been a problem for Intel's mobile system-on-chip teams in the past (Broxton, for example, was supposed to be a "mid-2015" launch), and I don't think any prudent investor should take it completely on faith that Intel will deliver this time.

However, given the strength of its data center business, and given that its PC business continues to show signs of stabilization, I'm willing to wait to see how 2016 plays out for the company's mobile efforts.