Has Intel Lost Its Manufacturing Lead?

The crown jewel of Intel's (NASDAQ: INTC  ) competitive advantage is, above all else, the company's leading-edge manufacturing technology. In today's world, any company can license great CPU and graphics IP and develop a processor for just about any market segment other than PCs. The only way Intel can truly differentiate is by having manufacturing technology that no other company can match. While Intel claims to have a manufacturing lead, it's worth peeling back these claims and understanding exactly what's going on here.

OK, what's everybody claiming?
Intel is claiming the following in terms of product rollouts:

  1. Cherry Trail (14-nanometer tablet product) launching in late 2014.
  2. Broadwell (14-nanometer PC products) launching in H2 2014.
  3. Broxton (14-nanometer smartphone/tablet converged product) launching in H2 2015.
  4. SoFIA 14-nanomater (14-nanometer low-cost smartphone/tablet product) launching in either Q4 2015 or Q1 2016.

On the other hand, Taiwan Semiconductor (NYSE: TSM  ) and Samsung, the two remaining leading-edge foundries, are claiming that their 14/16-nanometer process technologies will be in high volume production during 2015. If we look at what has been announced so far in the way of products and timelines, we can get a sense of Intel's competitive position:

  1. Qualcomm (NASDAQ: QCOM  ) has announced 20-nanometer Snapdragon parts for sampling in H2 2014 and in devices by H1 2015.
  2. Apple (NASDAQ: AAPL  ) is reputed to be TSMC's major 20-nanometer customer for an iPhone 6 launch in the August-September timeframe.
  3. AMD (NYSE: AMD  ) has indicated that it will be moving to 20-nanometer during 2015 and then it will roll out designs based on 14/16 FinFET at some point during 2016 (likely mid- to late 2016 if the current product release cadence holds).

Let's focus on Qualcomm
Intel's fiercest direct competitor in the chip space is Qualcomm, so it's worth taking a look at what the product release cadence for Qualcomm's products on new manufacturing technologies has been. (These are first device launches, so silicon is available a few months beforehand.)

  • April 2012 -- 28-nanometer polysilicon (Snapdragon S4).
  • July 2013 -- 28-nanometer high-K/metal gate (Snapdragon 800).
  • H1 2015 -- 20-nanometer high-K/metal gate (Snapdragon 808/810).

The gap between 28 polysilicon and 28 high-k/metal gate was 15 months, and the gap between 28-nanometer high-k/metal gate and 20-nanometer high-k/metal gate (assuming April 2015 product availability of Snapdragon 808/810) looks to be around 20 months. This has some interesting implications.

Case No. 1: 20 -> 16 transition is similar to 28 polysilicon to 28 HKMG
TSMC claims that the 16 FinFET process is simply introducing faster transistors onto a similar 20-nanometer back end of line, which should theoretically allow for a much shorter time to market for a 16 FinFET part than if the 16 FinFET node were all-new. If we assume, then, that the transition from 20-SoC to 16-FinFET is then similar to the 28-polysilicon to 28-high-K/metal gate transition, this would imply a 12- to 15-month gap, landing a Snapdragon 808/810 FinFET based successor in devices by Q2-Q3 2016.

Case No. 2: 20->16 transition is similar to 28 HKMG -> 20 SoC
Since a move to 16 FinFET is a move to a fundamentally new transistor structure, there may be complications on the design side in getting these products to yield enough for high volume. This, then, may make the 20 -> 16 transition look a lot more like the 28 HKMG -> 20 SoC transition in terms of time rather than the 28 polysilicon -> 28 HKMG. If that's the case, then the first 16 FinFET products from Qualcomm become a late 2016/early 2017 affair -- which is, not coincidentally, what International Business Times' Handel Jones appears to believe.

What does this mean for Intel?
Intel saw the first high-volume mobile SoCs based on its 22-nanometer FinFET manufacturing process appear in the September-October 2013 timeframe, with more aggressive volumes ramping throughout 2014. If Qualcomm gets 16 FinFET products out in devices by Q2 or Q3 2016, then Intel will have a lead of about a year and a half (Cherry Trail will probably appear in devices in Q1 2015) in tablets and about a three- to four-quarter lead in smartphones. This is a real lead, but it is not on the order of years as Intel has promoted.

Now, if Qualcomm and the foundry landscape slip to much later (something that seems doubtful), then Intel's lead in manufacturing becomes much more pronounced and the competitive environment becomes a lot easier for Intel. Further, if Intel truly has a density lead on the order of a 35% shrink relative to TSMC/Samsung 16/14-nanometer, then Intel's lead is quite obvious and -- assuming Intel's design groups don't fumble it -- should translate into a good competitive position for the company in late 2015 and beyond.

This chip stock could make you rich!
If you thought the iPod, the iPhone, and the iPad were amazing, just wait until you see this. One hundred of Apple's top engineers are busy building one in a secret lab. And an ABI Research report predicts 485 million of them could be sold over the next decade. But you can invest in it right now, for just a fraction of the price of Apple stock. Click here to get the full story in this eye-opening new report.


Read/Post Comments (6) | Recommend This Article (3)

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 Report this Comment icon found on every comment.

  • Report this Comment On May 13, 2014, at 8:58 AM, SSchlesinger wrote:

    You never mentioned the cost of such products. 20nm for TSMC is in a best case scenario will be on the same cost basis as 28nm to more expensive than 28nm. 16nm by the slides that Nvidia showed in 2012 and has since been shown around by other vendors will be at a significantly higher cost per wafer, perhaps double. Can QCOM show up by your account a year and half later, perhaps longer and offer a product in a market that will be significantly higher than Intel? Are they ready to go the contra revenue route like Intel has gone?

  • Report this Comment On May 13, 2014, at 9:51 AM, TMFAeassa wrote:

    @ SSchlesinger

    QCOM has more money than Intel does in the bank and generates roughly similar net income annually. I would not underestimate QCOM's purchasing power at the fab.

  • Report this Comment On May 13, 2014, at 10:03 AM, raghu78 wrote:

    Ashraf

    Samsung and TSMC expect 14/16 FINFET volume manufacturing to start by early 2015. If we are pessimistic they might start late q1/ early q2 2015. TSMC 16 FF is yielding much better than 20 SOC at the same point in its life cycle. So no major show delays are expected.

    https://markets.jpmorgan.com/research/email/-kjegkq4/GPS-133...

    http://www.tsmc.com/uploadfile/ir/quarterly/2013/4mMqe/E/TSM...

    "Then I cover the updates on 16 FinFET, 16 FinFET plus and our 10 FinFET. First, we have two general offers for customers, 16 FinFET and 16 FinFET plus. 16 FinFET plus offers 15% speed improvement, the same total power, compared to 16 FinFET. More importantly, 16 FinFET plus offers 30% total power reduction at the same speed, compared to 16 FinFET.

    Our 16 FinFET plus matches the highest performance among all available 16-nanometer and 14-nanometer technologies in the market today.

    Compared to our own 20 SoC, 16 FinFET plus offers 40% speed improvement. The design rules of 16 FinFET and 16 FinFET plus are the same; IPs are compatible. We will receive our first customer product tapeout this month. About 15 products planned for 2014, another about 45 in 2015. Volume production is planned in 2015. Since 95% tools of 16 and 20 are common, we will ramp them in the same gigafabs in TSMC. 16 FinFET yield learning curve is very steep today and has already caught up with 20 SoC. This is a unique advantage in TSMC 16-nanometer."

    Apple and Qualcomm are driving leading edge process node demand and are also multisourcing their designs. There is a very high probability that A9 is manufactured at TSMC and Samsung / GF.

    http://www.digitimes.com/news/a20131218PD207.html

    "Samsung will use its 14nm FinFET process to manufacture a portion of the A-series chips for Apple's 2015 iPhone series, and the remainder of the chips will be built using TSMC's 16nm FinFET technology, the sources indicated. TSMC should land 60-70% of Apple's total 14/16nm chip orders, with the remaining 30-40% to be taken over by Samsung, the sources said."

    First tapeout at TSMC 16FF+ happened last month. Given that A9 has to be in HVM by Q2 2015 for a late Q3 / early Q4 2015 launch I am guessing that the first tapeout at 16FF+ is Apple's A9.

    There is no significant density advantage to Intel as TSMC 16FF+ and Samsung 14 LPE /14LPP provide 15% area scaling from 20nm hkmg planar which in turn provides a 1.9x density scaling over 28 nm hkmg planar. TSMC 16FF+ matches Intel 14nm in transistor performance. Intel will face fierce competition at leading edge going forward. As Intel has big mobile ambitions, the ARM ecosystem is pushing the foundries to close the process gap to better compete with Intel.

    The foundries closing the gap at leading edge nodes also helps AMD (new high performance x86-64 core and custom ARMV8 core (K12) based servers) and Nvidia (custom ARMv8 Denver core based servers) and other ARMv8 licensees like Applied Micro, Marvell with their server efforts. ARM has put a lot of effort into providing the various building blocks for ARMv8 servers. - cores, coherent fabric (with plans for chip to chip coherent interconnect planned like AMD Hypertransport or Intel's QuickPath )

    http://techreport.com/review/26418/amd-reveals-k12-new-arm-a...

    http://techreport.com/review/26427/arm-lays-the-foundation-f...

    http://www.fudzilla.com/home/item/34705-nvidia-might-enter-a...

    All these developments lead to an epic battle between Intel and ARM ecosystem and this is something I look forward to.

  • Report this Comment On May 13, 2014, at 10:07 AM, SSchlesinger wrote:

    @ TMFAeassa

    Let me get this straight because I am having a hard time believing this. QCOM is going to have have an 18 month delay at the minimum, which is a full generation in the computer industry and catch up? They risk losing design win after design win by this point to Intel who will be moving to 450mm wafers around this time AND will be working on 10nm.

    As much as I respect Ashraf's analysis you guys are both out on a limb on this one.

  • Report this Comment On May 13, 2014, at 11:11 AM, TMFAeassa wrote:

    @ SSchlesinger

    In tablets, Intel has about an 18 month lead in process technology, yes. However, in the much higher volume/dollar content smartphone market, the lead is appreciably narrower -- about 3-4 quarters.

    If Intel can leverage that lead to win major designs and then build off of that momentum, then Intel is in a good position. If we end up with delays and products that don't quite hit the mark, then 14-nanometers or not, Intel still doesn't win the business.

    Time will tell how this all plays out.

  • Report this Comment On May 13, 2014, at 3:53 PM, ethos wrote:

    I don't understand why you title this article with something that has a negative implication when you could just as easily have said, "Intel maintains a manufacturing lead over its rivals". Do you wish to create fear and doubt? I think a completely fair and honest person would have titled this article differently.

Add your comment.

DocumentId: 2954491, ~/Articles/ArticleHandler.aspx, 8/1/2014 6:42:10 AM

Report This Comment

Use this area to report a comment that you believe is in violation of the community guidelines. Our team will review the entry and take any appropriate action.

Sending report...


Advertisement