Over on micro-blogging website Weibo, one poster claims that there are two variants of LG's next-generation applications processor, known as the NUCLUN 2, in the works. The first is built by Taiwan Semiconductor Manufacturing Co. (TSM 3.16%) on its 16-nanometer FinFET Plus process. The second is built by Intel (INTC 0.55%) on its 14-nanometer manufacturing process.
According to the post, the version built by Taiwan Semi runs its "big" ARM Cortex A72 cores at 2.1 GHz while the one built by Intel is able to run those same cores at 2.4 GHz (though a red flag should go up right now because HiSilicon recently announced that its Kirin 950 chip, built on TSMC's 16-nanometer FinFET Plus process, features Cortex A72 cores that run at up to 2.3 GHz).
All told, the poster claims that the Intel-built version is able to deliver significantly better performance in the popular Geekbench 3 benchmark (15% better single-core score, 17.1% better multicore score).
Although I am hesitant to take this as gospel, I think there are a number of interesting implications here if this report is accurate. Let's take a closer look.
Intel may be attacking the high end of the phone market differently
Intel had previously tried to go after the high end of the smartphone market with home-grown applications processors. This strategy so far has not really worked out well for the company for a variety of reasons.
The first is -- to be perfectly blunt -- that the company simply has never put out competitive products. Intel's chips in this space have typically lagged in CPU/GPU performance relative to chips from its competition (i.e. Qualcomm (QCOM 0.17%) and, to a lesser extent, MediaTek) and have often trailed in other key functionality such as video encoding/decoding and imaging.
It also doesn't help that Intel's competitors have been integrating leading-edge LTE modems into their chips while high-end Intel platforms have required to be paired with stand-alone modems.
Beyond that, though, Intel faces the challenge of a relatively small served addressable market, or SAM, within the high end of the market. Many of the major players here have expressed an interest in developing their own processors, limiting the opportunity for merchant vendors.
And, for those customers that cannot afford to design their own chips, Qualcomm generally provides best-in-class solutions (in most cases better than the home-grown attempts we've seen from some device vendors) and the company has generally proven itself as reliable supplier in this market.
Although Intel may very well continue to try to develop merchant solutions for this market (though if the company is, its executives certainly didn't seem interested in talking about them at the company's recent investor meeting), manufacturing chips for those who want to design their own may be a viable strategy.
Not a game changer, but earning customer trust would be a good start
If Intel does actually wind up building LG's next-generation applications processor, then investors shouldn't bank on such a development being a financial game changer. LG will probably just deploy it in select high-end handsets in certain regions, with Qualcomm Snapdragon processors likely powering the majority of LG's high-end volume for the foreseeable future.
It's also worth noting that although LG is a fairly well-known name in the smartphone market, it's not likely that its share of the high end of the market (which is essentially dominated by the iEverything maker and, to a lesser extent, Samsung) is high enough to make such a win all that financially exciting.
Although such a win wouldn't mean big bucks for Intel, it might indicate that Intel has a higher-performing process for mobile applications (as the report claims) and is becoming better as a foundry partner, something that Taiwan Semi is widely regarded as the best in the business at.
There's far more to the equation here than the underlying technology -- a point that former Intel Custom Foundry chief Sunit Rikhi made in an interview for online publication Semiconductor Engineering -- so solid progress with a customer like LG might make it easier for the company to win other high-end mobile foundry business over time.
