While semiconductor manufacturing technology isn't the only thing that determines the competitiveness of a given chip, it is one of the key enablers of dramatic performance and power improvements. Intel (INTC 5.69%), which has struggled mightily to develop chips that mobile device vendors actually want to use, is well known for its manufacturing technology lead over the rest of the industry. This lead isn't just in the ability to manufacture the latest-generation parts, but it's also in that Intel's design and manufacturing teams are very closely aligned.
Some background on 14-nanometer
Today, Intel ships chips built on the 22-nanometer FinFET manufacturing process, while its competitors ship on 28-nanometer planar. FinFETs essentially allow a transistor to deliver significantly more performance at lower power than traditional "planar" devices of the same geometry. This, in theory, should have given Intel a pretty significant advantage at the 22-nanometer generation over its 28-nanometer counterparts, but of course Intel's competition had superior chip designs, compensating for the manufacturing deficit.
However, it is likely that Intel will significantly improve the competitiveness of its designs at the 14-nanometer node with Cherry Trail (for tablets) and Broxton (for higher-end tablets and high end smartphones). Further, Intel claims that its 14-nanometer process yields significant density improvements over Samsung's (NASDAQOTH: SSNLF) 14-nanometer process and Taiwan Semiconductor's (TSM 4.96%) 16-nanometer process, which should allow Intel to pack more functionality into a given area than its competitors, leading to product superiority -- in theory.
Let's talk product timelines
Intel plans to launch its first 14-nanometer tablet processors in late 2014 (think November), and its first smartphone 14-nanometer product by "mid-2015" (let's say that this is the July-August timeframe). Qualcomm (QCOM 2.19%), Intel's biggest rival in the mobile space, will have 20-nanometer planar-based product in the marketplace during the first half of 2015 (with silicon probably shipping to partners in early 2015 for April-May 2015 device availability). It stands to reason that Qualcomm will have 14/16 nanometer product available within a year of that.
If Qualcomm has 14/16-nanometer product in the first half of 2016, then this gives Qualcomm and Intel a period of "overlap" (although again, Intel claims its 14-nanometer process is superior) before Intel presumably transitions to its first 10-nanometer Atom products in mid-2016. At this point, Intel will be on the 10-nanometer node (which should be 1.5 to two generations ahead in density relative to TSMC and Samsung-built 14/16-nanometer), which means that Intel will have the largest manufacturing lead it has ever had in the mobile space on both transistor performance and transistor density.
If this comes true, then Intel could do quite well
While the semiconductor manufacturing technology isn't the be-all, end-all of product performance, as long as Intel's product teams do a good job defining and executing next-generation products, Intel should have a clear power and performance lead over its competitors by mid-2016. While this doesn't solve Intel's Android software compatibility issues, the hardware should be unequivocally leadership, if it wasn't already at the 14-nanoeter generation.
Of course, there's plenty of risk here. We saw that even the mighty Intel struggled to ramp 14-nanometer to the right yield levels, so there is risk that the 10-nanometer node may be delayed. Further, even if Intel can ramp 10-nanometer to production yields, the designs need to be there from both a hardware and software perspective. And if all of that is there, Intel still needs to fight Qualcomm (as well as potentially in-house chip designs at some of its customers) for designs, which is a battle that goes beyond technical specifications.
Foolish bottom line
At the 10-nanometer generation, Intel is likely to have a pretty significant performance, watt, and cost advantage over its peers. This has a good chance of translating into unequivocal product leadership, although it is important to understand that Intel's product definition and implementation teams need to execute properly for this manufacturing advantage to translate into a real-world sales advantage.
