Intel's (NASDAQ: INTC ) competitive position in the ultra-mobile market continues to improve with each generation. At the 32-nanometer generation with Medfield and Clover Trail, Intel proved that there were no fundamental barriers keeping Intel from penetrating the mobile market (i.e., busting the X86-myth). At the 22-nanometer generation, Intel's parts in both tablets and phones were solidly competitive, particularly on the CPU side of things. The 14-nanometer generation, which should arrive next year, is expected to be the one that "seals the deal."
Intel's 14-nanometer: It's Broxton that matters
By the very end of 2014, Intel expects to launch its very first 14-nanometer mobile processor code-named Cherry Trail. According to Intel's Cara Walker, this part will be aimed at performance and mainstream tablets -- including 2-in-1s and "phablets" -- although it's unclear if Intel's definition of "phablet" is more along the lines of a 7" voice-enabled tablet, or a 5-6" handset.
However, while the Cherry Trail parts should do well in the tablet market, Intel's big hurdle -- and the market that is by far the more important from a growth perspective -- has been the handset market. According to Intel's Investor Meeting, Intel's first truly high-end, "hero device" worthy system-on-chip for phones and tablets is called "Broxton." It will be based on the Goldmont CPU architecture and Intel's next-generation Gen. 9 GPU. It is also targeted at a "mid-2015" launch.
Source: Intel.
Who are the chief competitors?
Of course, no competitive analysis is complete without a discussion of what the competition will be doing. As of today, Intel's chief competitors in the high-end handset apps processor space are the following:
Samsung is particularly potent given that it also sells about 30% of all smartphones globally and could simply choose to block Intel from competing for that large chunk of the market. Qualcomm is the most aggressive on the chip/integration side of things, and MediaTek is able to very quickly deliver solutions based on off-the-shelf IP.
Fast-forwarding to mid-2015
By the time mid-2015 rolls around, the competitive landscape (in terms of shipping process node from each of these players) should be the following:
- Qualcomm's MDM9x35 modem built on the 20-nanometer process will ship during the second half of 2014, with a 20-nanometer apps processor with integrated modem likely shipping to customers by Q4 2014 for late Q1/early Q2 2015 device launches.
- Samsung will likely have a 20-nanometer Exynos part (likely based on ARM's Cortex A57) ready for the Galaxy Note 4 launch in the September/October timeframe, if not for the Galaxy S6 in early Q2.
- MediaTek, due to its position as a low-cost/high volume player probably will not move to the 20-nanometer node until 2015 after it has matured and costs have come down.
So, this essentially leaves Intel competing with its 22-nanometer Silvermont-based, quad-core Moorefield in phones during the second half of 2014 until the Broxton launch in "mid-2015."
Intel's 2014 mobile platforms. Source: Intel.
But what does "mid-2015" mean?
According to Walker, the timelines given at the investor meeting were silicon ship/availability dates. So, "mid-2015" implies silicon availability in the June-August 2015 timeframe. Given that Merrifield was launched in February 2014 and won't appear in devices until May/June, it is likely that Broxton-powered phones won't appear until 3-4 months after silicon launch. This implies a September-November 2015 timeframe.
Intel's Hermann Eul demonstrating Merrifield. Source: Intel.
Realistically, TSMC and possibly Samsung could be in a position to offer top-paying customers 14/16nm FinFET wafers (even if the yields aren't great). For example, the Galaxy Note 5 and the iPhone 6s should both presumably launch in the Sept. 2015 timeframe and are great candidates for 14/16nm FinFET silicon if it is ready but still expensive. Of course, Intel's 14-nanometer process is denser and likely higher performing, but the gap between Broxton and potential Apple/Samsung/Qualcomm FinFET chip implementations may not be as wide as the process availability gap would imply.
So, what's the bottom line?
If TSMC/Samsung can get 16/14nm FinFET designs into shipping devices by Sept. 2015, then Intel's manufacturing lead will have shrunken to less than one generation. If they cannot and those designs are using 20-nanometer silicon, then Intel's lead will actually be quite wide. Also, if Intel can advance to 10-nanometer designs quickly after the 14-nanometer Broxton (i.e., by mid-2016), then that could be the real "winner" while Broxton is merely "very competitive."
Tired of waiting for Intel to deliver in mobile?
Let's face it, every investor wants to get in on revolutionary ideas before they hit it big. Like buying PC-maker Dell in the late 1980s, before the consumer computing boom. Or purchasing stock in e-commerce pioneer Amazon.com in the late 1990s, when it was nothing more than an upstart online bookstore. The problem is, most investors don't understand the key to investing in hyper-growth markets. The real trick is to find a small-cap "pure-play" and then watch as it grows in EXPLOSIVE lockstep with its industry. Our expert team of equity analysts has identified one stock that's poised to produce rocket-ship returns with the next $14.4 TRILLION industry. Click here to get the full story in this eye-opening report.
RSS Headlines
Fool UK
Comments from our Foolish Readers
Help us keep this a respectfully Foolish area! This is a place for our readers to discuss, debate, and learn more about the Foolish investing topic you read about above. Help us keep it clean and safe. If you believe a comment is abusive or otherwise violates our Fool's Rules, please report it via the
Report this Comment icon found on every comment.
Report this Comment On April 02, 2014, at 12:47 AM, SSchlesinger wrote:
"If TSMC/Samsung can get 16/14nm FinFET designs into shipping devices by Sept. 2015, then Intel's manufacturing lead will have shrunken to less than one generation."
Is this an April Fool's joke? These ARM timeframes seem a bit optimistic. Samsung will go past 20nm planar to 14nm by the middle of next year and figure out finfet along the way. They will somehow be able to go through the yield issues and have this all figured out in 18 months? Intel has struggled getting 14nm working and they've had 22nm for a couple of years. BTW, are you also projecting that they will be at 450mm wafers like Intel will be at the time?
Report this Comment On April 02, 2014, at 12:55 AM, TMFAeassa wrote:
@ SSchlesinger,
Thanks for the comment. Notice my liberal use of the word "if" :-)
At the end of the day, we'll have to see. TSMC claims that it'll be in volume production on "16 FinFET" by the end of this year, but let's call it about a year after the "volume production" begins for product to show up in meaningful quantity. This could mean late 2015/early 2016 for the most aggressive SoC companies...if TSMC can deliver.
Samsung talks a big game, but TSMC's R&D and semiconductor revenues are much higher. The rumors around claim that Samsung's 20nm is having severe problems, implying that TSMC is ahead here.
We'll have to see how it all plays out. Intel will be first to 14nm SoCs with "Cherry Trail", but in phones the timeline for Intel v.s. the top ARM guys is less aggressive.
Report this Comment On April 03, 2014, at 3:43 PM, raghu78 wrote:
" Of course, Intel's 14-nanometer process is denser and likely higher performing, but the gap between Broxton and potential Apple/Samsung/Qualcomm FinFET chip implementations may not be as wide as the process availability gap would imply."
This JP Morgan report has positive remarks on state of TSMC 20nm / TSMC 16FF / TSMC 16FF+
https://markets.jpmorgan.com/research/email/-kjegkq4/GPS-133...
Basically TSMC 20nm is in volume production with a very steep ramp in H2 2014. 10% of 2014 and 205 of Q4 2014 wafer revenue at 20nm according to TSMC mgmt at Q4 2013 earnings call.
TSMC 16FF in volume production by late Q4 2014. TSMC 16FF+ behind by 1 -2 quarters. Mobile customers might go directly to TSMC 16FF+.
TSMC 16FF+ matches Intel 14nm in transistor performance. This would be the first time ever for TSMC that they have achieved transistor performance parity. TSMC 16FF+ is closely matched in transistor density though slightly behind. There is no significant density advantage to Intel.
Intel's job just got harder. Apple is going to make any revenue growth for Intel almost impossible. the A8 and A9 will continue to cannibalize the x86 notebook , particularly enterprise with the availability of Office for ipad. Qualcomm also should have an ARM V8 custom core ready by 2015 end. Intel's challenge would be to avoid revenue reductions going forward.
Add your comment.