Investors following Intel's (NASDAQ:INTC) mobile plans know that the company is planning to introduce a lineup of chips that it refers to as SoFIA. During 2015, Intel hopes to launch a series of such products targeting both 3G phones as well as LTE phones. The funny thing about these initial products is that they will be built on Taiwan Semiconductor's (NYSE:TSM) 28-nanometer process, not Intel's vaunted 22-nanometer or 14-nanometer processes.
From Intel's public statements, this is being done for time to market reasons. Intel plans to bring 14-nanometer SoFIA products to market in the 2016 timeframe (which should help Intel's competitive positioning and economics), but investors may wonder why Intel is bothering with the initial 28-nanometer SoFIA products at all.
Why not just wait until 2016?
Intel claims that its SoFIA products will be competitive, but until we see actual performance tests from third parties, it's hard to make any definitive statements. I will say though that in the low-end and mid-range, I fully expect Intel's 14-nanometer SoFIA products to be much more competitive on a relative basis than the 28-nanometer ones since Intel will likely have an actual manufacturing edge there.
At any rate, so this brings us to the question of "why bother?" If Intel didn't bother building these products and instead dedicated the R&D money and staff to bringing the 14-nanometer SoFIA products out a bit sooner (or deployed them to the 10-nanometer SoFIA parts), then wouldn't this be a better use of capital?
Well, it's not so simple, and there are two major reasons that I can think of that Intel is likely developing these 28-nanometer SoFIA products.
Fixing the bill of materials issue
As investors may be well aware, the company's 22-nanometer product family (Bay Trail, Merrifield, and Moorefield) suffer from a bill of materials problem. This means that if a tablet vendor wanted to go ahead and build a tablet around an Intel platform, it would need to spend likely $15 or more extra per tablet (per CFO Stacy Smith at the 2014 investor meeting) than it would with an equivalent ARM (NASDAQ:ARMH) based platform.
I suspect part of this "bill of materials bloat" is due to the fact that Intel's chips aren't as integrated as the competition's, but the company has also talked about a lot of other non-SoC related costs relating to the actual platform design.
Intel was able to bring that delta down somewhat with a platform known as Bay Trail-CR, but the company has said that it won't be until SoFIA (and Broxton) hits the market that this disadvantage goes away.
If Intel didn't do the 28-nanometer SoFIA parts and waited for 14-nanometer SoFIA, then Intel would likely have needed to provide more contra-revenue offsets this year as it would have needed to use the 22-nanometer product lineup to achieve this goal -- hurting revenue and profits.
Another thing that Intel needs to do is gain credibility in the mobile market. While Intel was able to do so in tablets with Bay Trail (coupled with contra-revenue dollars), it may have been prohibitively expensive to provide such contra-revenue in low-end phones in addition to tablets during 2015.
So, if Intel wants to gain credibility in the low-end of the smartphone market -- something that Intel has made clear that it wants to do, it needs to deliver competent products on time. If Intel can do this with the first generation of SoFIA with a handful of customers, then by the time the 14-nanometer SoFIAs come in, Intel not only has a set of repeat customers from the first SoFIA wave (mirroring what Intel did in tablets with Bay Trail), but it could also attract new customers who have been on the fence about using Intel platforms in smartphones.