Intel (INTC +1.15%), the world's largest semiconductor company by revenue, has faced significant difficulty gaining material traction in the smartphone and tablet processor market. While the company plans to get the proverbial ball rolling on tablets this year with a goal to ship 40 million units during 2014, smartphones remain elusive as the company has yet to offer an integrated apps processor and leading-edge modem solution.
Intel's SoFIA stop-gap
At its 2013 Investor Meeting, Intel announced that it would be launching a product family known as SoFIA. The first instantiation, coming in late 2014 likely for low-end handsets launching in early 2015, is known as SoFIA 3G, which features two Intel-designed processor cores coupled with a 3G modem built on TSMC (NYSE: TSM) 28-nanometer process. In early 2015 (likely Q1), Intel will launch a quad-core version of this part with an integrated LTE modem.
These parts should remedy Intel's biggest weakness in the handset market: lack of integrated comms. While high-end phones can support the increased bill-of-materials associated with less-integrated applications processors, the low end is exceptionally cost sensitive as the competition is fierce and the barriers to entry low. A difference of $5-$10 on the bill of materials for a phone that sells for under $200 is material, and in this segment of the market, additional performance isn't as important as keeping costs down. Intel's SoFIA products integrate comms and connectivity, keeping platform costs very low.
Intel needs to move this internally ASAP
Intel's plans have always called for the successor to the original TSMC-built SoFIA product to be built internally on Intel's 14-nanometer process. This should give roughly two generations' worth of area scaling, higher performance transistors, all coupled with what is likely a higher-performance design. At Intel's Investor Meeting, CEO Brian Krzanich indicated that by the end of 2015, the 14-nanometer SoFIA would come to market (likely for Mobile World Congress 2016 device launches).
However, on the most recent call, Mr. Krzanich had the following to say with respect to this next-generation, internally built SoFIA:
We'll bring that in on our 14-nanometer process either late 2015 or early 2016. We're still battling back and forth on how fast we can bring it in and what impacts that has. 14-nanometer is the technology there.
If Intel brings that product to market in late 2015, then the product could make it for device launches in the Mobile World Congress 2016 time frame. However, if the chip is launched at Mobile World Congress 2016, then it won't appear in devices until about mid-2016. This would significantly dull any competitive edge Intel may have had with this product, although it would still be quite competitive for the entry/value segments at which it is targeted when it is finally released.
It's not just about competitive edge
While moving SoFIA internally as quickly as possible gets Intel a better cost structure and better performance, the real problem here is that TSMC will not only profit from products built by Intel's direct chip design competitors, but it will profit from Intel's own chip sales. Helping TSMC profit is the last thing Intel wants, so the sooner TSMC stops getting Intel's wafer orders, the better it will be for Intel, especially as Intel will be able to drive gross margins up by driving factory utilization up. The low cost smartphone market is a very high-volume market, and it is exactly these kinds of volumes Intel should be aggressively targeting for its own factories.
Foolish bottom line
At the end of the day, Intel needs to get a big chunk of that mobile chip goodness not only to bolster its own business results, but to keep as much ammo out of the hands of the rest of its competitors as possible. In just a few short years, Qualcomm (NASDAQ: QCOM) has gone from a phone chip vendor to a leading semiconductor powerhouse that is now TSMC's largest customer.
The time for baby steps and half-efforts is over. Intel needs to build leadership mobile parts on its own manufacturing processes within Intel's factories, and it needs to do so as quickly as possible. Intel can still be a major player in the mobile silicon market, but that window of opportunity shrinks each and every day as its competitors continue to advance their technology and designs at breakneck speeds.
