With 2013 performance of only 10 million tablet chip sets (for Windows mostly) Intel is still confident in its ability to deliver 40 million of those (with increased Android portion) in 2014. To achieve this they will be doing a lot of enabling across the industry to take the Bay Trail-based tablet BOM cost down to an equivalent level. They expect that the company’s overall margin will be hit just by 1.5% because of this required in 2014 effort. They are saying that Intel will be safe from 2015 on as moving to 14nm process technology with next-generation (even in terms of micro-architecture) Broxton and SOFIA SoCs for tablet and smartphone devices. They are basing this statement on their inherent “transistor density” advantage against TSMC from that point in time on, despite some analysts’ opinion of the economy of scale advantage of TSMC in terms of the number of wafers produced.
With media generally reporting that Acer’s biggest mistake was its too early and too heavy bet on ultrabooks it is clear that OEMs will take a very cautious approach with Intel’s efforts to decrease the Bay-Trail based tablet costs down on the BOM level, as it is exactly what happened with ultrabooks. Instead the will try to solidify their tablet market position with ARM-based tablets in all segments of the tablet market, from the lowest cost upto the premium. Moreover, Jason Chen’s appointment to the CEO position of Acer is also showing that even for ongoing efforts OEMs need a very detailed and deep understanding of the SoC manufacturing and even the process technologies. Take note of Jason Chen’s history of employment in order to understand that:
- TSMC: 2005-2013
- Intel: 1991-2005
- IBM: 1991-1998
In other regards we only know that Acer to start new operation strategy in April to focus on BYOC (Build Your Own Cloud) [DIGITIMES, Jan 13, 2014] and that “In the future, all of Acer’s businesses including desktop, notebook and tablet will involve the BYOC platform and it is hoping to strengthen its product lines through the services.” It will be interesting to watch what that means as my previous conclusion was Leading PC vendors of the past: Go enterprise or die! [‘Experiencing the Cloud’, Nov 7, 2013].
Now back to the Intel related information in terms of details in their earnings call. Note before that the correlation of Intel and Microsoft stock prices (as well that the stock market was absolutely not happy with Intel results and especially with the “flat 2014” outlook):
The company’s stance for 2014 is indeed not rosy as Intel to reduce global workforce by five percent in 2014 [Reuters, Jan 17, 2014].
From: Intel’s CEO Discusses Q4 2013 Results – Earnings Call Transcript [Seeking Alpha, Jan 16, 2014]
Inserted slides are from Investor Meeting – Stacy Smith (CFO) [Nov 21, 2013] while the acompanying text is from Intel Shares Mobile Progress, Priorities and Product Pipeline at Annual Investor Day [Technology@Intel, Nov 25, 2013] if reference is not put underneath
[On transistor density and wafer cost]
Mark Lipacis – Jefferies
Thanks for taking my question. At the Analyst Day, you addressed your view on transistor density and your expectation for leadership on that vector, but I have to say this discussing that idea with investors is a consensus view that seems to be that Intel has an inherent wafer cost disadvantage that relative to TSMC that neutralizes or more than neutralizes your transistor density advantage and the argument is that TSMC ships more wafers and therefore has more better purchasing power than you and its lower labor cost, so net-net, they have just a big huge advantage of wafer cost that you should have a hard to, too hard of a time to overcome. So my question is do you think that’s a fair view. Can you help us talk to the relative elements of the wafer cost and how you think you can compare? Any kind of help that you give us on the cost dimension would be extremely helpful. Thank you.
From: CES: Process Will Still Win in Mobile, Says Intel’s Eul [Barrons.com, Jan 9, 2014]
Eul points out that Qualcomm, and other competitors such as Nvidia (NVDA) and Broadcom (BRCM), all of whom are dependent on Taiwan Semiconductor Manufacturing Company to actually make the chips they design, will run into a problem as Taiwan Semi’s technology stops scaling.
Intel had made the point at the analyst day presentation, and Eul repeated it: As TSMC moves from 28 nanometer to 20 nanometer, it will run into a problem at the subsequent step, 16 nanometer, where TSMC will not add any real reduction in transistor size. That, says Eul, means that 16-nanometer parts a few years from now will be stuck at a 20-nanometer feature size while intel presumably zooms ahead to 10 nanometer by that time.
And what that means is that, unable to scale the density of a chip as Intel can, Qualcomm and Nvidia and Broadcom and the others will not be able to integrate as many parts as Intel on a single semiconductor die.
And so to those who point out that Intel hasn’t yet released its integrated baseband chip, Sofia, mentioned above, Eul contends the company will have the last laugh in a few years’ time as Qualcomm and the rest hitting a scaling wall.
Brian Krzanich – Chief Executive Officer
You know I think the first thing to remember is that what really counts in all of this is transistor cost and what we really talk about in our Moore’s Law of Curves and when we talk about transistor density is driving a consistent cost reduction of the transistors and so wafer cost is one segment of that. I’m not going to comment on you know TSMC’s wafer cost versus our wafer cost but we feel confident that our relative level of scaling and our internal wafer cost are such that we believe we have a leadership position in transistor cost.
When you’re talking about any product whatever it is, a logic product that’s a low-end microprocessor for wearable or internet of things or high-end Xeon server. You’re talking about the number of case and hence the number of transistors required to put that logic device together, it doesn’t matter whose technology it’s on to some extent. It doesn’t matter what node and so the more cost effective those transistors are whether it’s 500 million or 3 billion the lower the product cost there is and that’s really what we focus on and why we focus on transistor cost. So I think we stand by our what we said at the investor meeting.
Brian Krzanich: Our disclosure in November of a new smartphone and tablet road map that will include SoFIA our first IA SSD with integrated comps later this year is further evident that we’re innovating and bringing products to market at faster pace. Looking ahead 2014 will be an exciting year as we build further on this new foundation. We have established a goal to grow our tablet volumes to more than 40 million units. Within an emphasis on the value segment. As we’re finishing 2013 with more than 10 million units and a strong book of design wins we’re off to a good start.
Stacy Smith: In the tablet market, we launched the Bay Trail SoC and have started to expand our footprint and market signature in this growing market.
The 4X Tablet Campaign: This year, Intel increased its focus on tablets with key design wins and the introduction of Bay Trail. Next year, Intel plans to increase tablet volumes by 4X! Eul signaled a rich pipeline of tablet and phablet design wins for Bay Trail including Android and Windows devices spanning price points from premium to sub $99 products from leading OEMs and the China tech ecosystem. He also said industry leading performance, competitive battery life, cost-reduced SOCs and unique features like 64 bit will help drive growth. Intel gave a first-time demo of the performance gains achieved with a 64 bit Bay Trail system running Windows and showed a 64 bit kernel running on an Android tablet.
Note the details about the 2014 tablet market of ~289+ million units in the 2014 will be the last year of making sufficient changes for Microsoft’s smartphone and tablet strategies, and those changes should be radical if the company wants to suceed with its devices and services strategy [‘Experiencing the Cloud’, Jan 17, 2014] post of mine. The 40 million target of Intel is therefore less than 14% of that.
[regarding: So on the tablet strategy to get the 40 million you’re saying it’s going to be a 1.5 percentage hit.
Gross Margin Reconciliation: 2013 to 2014 Outlook (59.8% to 60% +/- a few points)
– 1.5 points: Tablet impact
Let’s say you guys get into the second half of the year and you’re not quite to the 40 million if it’s a pretty significant short fall. Would you consider canning that strategy I guess I’m just wondering what the commitment is if the volumes aren’t there but the cost is there by the end of the year?]
Brian Krzanich: This isn’t a price reduction as normal price reduction would be; it’s not where you are just simply reducing. It’s truly a BOM cost equalizer and remember a lot of our 40 million tablets in ’14 will be based on Bay Trail. Bay Trail was originally designed for Avoton-based PC segments and the upper end tablet [and all Windows]. And so it’s what we are doing here is doing a BOM cast delta relative to the, what the mid and lower end tablets require. And so those are things like Bay Trail may require more layers of a printed circuit board for the board itself, more components on the board and tighter power management controls and things like that. We have a whole program to reduce those throughout the year. So that gives us confidence that as we go through the year, the BOM cast delta will shrink, but if the volume didn’t show up for some reason and I am not going to say that, that’s what’s going to happen, but I am confident it will, but if it didn’t it’s on a per unit basis. And so the spending on that contra would be reduced equivalently.
Stacy Smith: And I would just add as Brian said we are doing a lot of enabling across the industry to take the BOM cast out in equivalent. These are costs at the system level not at our chip level and it will vary a lot by SKU, but to give you a sense for a Bay Trail platform from the beginning of the year to the end of the year we think that, that BOM penalty drops by more than half. And so it kind of gets better out in time. And then when we get to the Broxton generation we think it’s de minimis.
Brian Krzanich: Both Broxton and SoFIA are just specifically designed to eliminate that delta.
Say “hello” to SoFIA: By the end of 2014, Intel will deliver a new integrated Atom processor + communications solution for entry and value smartphones and tablets, code-named SoFIA. In his presentation, Eul highlighted that Intel’s Infineon wireless assets make the company an “incumbent” in the mobile phone market, shipping more than 360M mobile platforms a year spanning 2G and 3G solutions. He said SoFIA builds on the proven 3G communications platform to deliver a competitive and highly integrated, IA-based mobile solution aimed at the fast-growing market for entry smartphones and tablets. The 3G version of SoFIA is expected by the end of 2014, and Eul said an LTE version would follow in the first half of 2015.
Accelerated Mobile Roadmap: While specific product details will be saved for a later date, Eul signaled a robust pipeline of new Atom processors and multi-comms solutions for 2014 and beyond to address devices spanning market segments from entry to performance smartphones and tablets, an approach he called “market-oriented pragmatism.” In addition to SoFIA, Eul noted:
Broxton – in 2015 Intel plans to deliver a 14nm, 64 bit SOC based on a new, next generation Atom architecture (Goldmont) targeted for hero devices. Broxton is being designed for pairing with Intel’s next generation LTE solutions.
[regarding: If we look at tablets and smartphone, what type of units do you need to reach for that business to stop having a material impact in gross margin from is 10 points higher utilization rates and excluding the contra revenue impact and that’s it? So just looking at the 40 million units target for this year, what type of volume do you need to get in order for gross margin to start appreciating from the west of the business if you exclude the contra revenue impact?]
Brian Krzanich: Yes, it’s hard to say. I mean, I will bridge back to our strategy here. Our strategy is that we are going to use our process technology leads. We will have leadership products that also are competitive or maybe even leadership in terms of cost and I showed some data at the investor meeting that just kind of showed the die size as we progress from Bay Trail to Broxton to SoFIA and so you can get a sense of the kinds of cost structure that we are going to have on a per unit basis. I don’t think it causes on a percentage basis. Yes, I can’t – I am not envisioning if this causes the gross margin percentage to go up, but you can definitely get to a space once we get through these contra enabling dollars where every unit we sell is accretive on a gross margin dollars per unit. It’s utilizing factories that we have in place for PCs. And so it’s a nice adder of that gross margin dollar per unit standpoint.
[regarding: Bay Trail Android tablets]
Brian Krzanich: Most of the Bay Trail Android tablets really start showing up more in Q2 than in Q1 and that’s again purely you know remember we made a shift, an original program for Bay Trail was all Windows. As we came into the midpoint of the year we sandbox [ph] shift and make it Windows and Android and so you know our OEM partners as well are targeting more towards Q2 and it’s just when you do you go and start putting back in that back to school event which is a next seasonal place where upside usually occur.
[regarding: On the smartphone or on tablet space, I think it is true that Intel has a manufacturing lead, but do you think your cost reduction efforts and then the Moore’s Law advantages ever progressed faster than the ASP declines in the space. In other words, do you think Intel can be sustainably profitable in the mobile space which is maturing?]
Brian Krzanich: Yes, we absolutely do. You saw at the investor meeting products like SoFIA, which really are going to be put on to 14-nanometer are fully integrated all the way through with the 3G option or an LTE option and that LTE is with carrier aggregation. Those kinds of products we believe are very, very cost competitive in fact leading from a cost position. In addition, we don’t talk a lot about, but we are already in that low cost Asia market. We are inch and then we are working with ODMs there. That’s actually where a lot of the innovations coming out of for some of these cost reductions on tablets and where we are getting the cost reduction ideas. So we are in that market now. We sold out of that Shenzhen low cost market in Q4. We will continue through it – through 2014 and with products like SoFIA on leading edge technology, we are very comfortable that we can get into those very low price points.