In this article, I'd like to give my thoughts/analysis on the recent earnings report and conference call from Intel (INTC).
PC Client Group: Not Bad, Encouraging Comments
In the most recent quarter, Intel's PC Client Group did $8.4B in sales - down 3.44% from the year-ago period. While a year-over-year decline isn't exactly the most encouraging trend to see in a company's bread-and-butter business, it is nice to see that the decline is slowing down. I believe that this decline is less dramatic than the PC numbers that IDC and Gartner have been giving (~7-9% decline) due primarily to the following drivers:
- The "netbook" category, which was actually a fairly high volume driver of PCs due to their low cost, isn't counted as part of Intel's "PC Client Group" revenues - it's part of "Other IA".
- Intel may be taking share from its nearest competitor, [[AMD]]), although this is not confirmed (nor did the company really mention it on the call)
However, while the numbers suggest that the PC demand environment is still declining, albeit in a less pronounced fashion than feared, management had some interesting comments on the call that I believe investors should at least take as a mildly encouraging sign,
The enterprise market for PCs strengthened in the third quarter and the consumer markets in the U.S. and Europe appeared to have bottomed out.
While this doesn't quite account for the fairly violent negative trends in the Asia-Pacific region (where tablets are replacing the lower cost netbooks that have now disappeared), this does suggest a couple of things:
- Even without the latest "Haswell" designs (2-in-1, convertibles, fanless, etc.), the PC market in developed countries still managed to find a bottom. Could this bode well for when the newer, more attractive designs hit the market?
- The bulk of the erosion really does seem to be at the lower end, and a non-trivial part of this is simply due to the fact that there are no attractive low end PC designs at this point that can fend off tablets. Could a combination of Intel-powered tablets, coupled with Bay Trail-M (low cost, fanless systems at low price points) drive adoption back at the low end, particularly in the Asia-Pacific region?
I'm not ready to become too optimistic on the PC market, but if the business can find a bottom relatively soon, then it becomes a whole lot easier to view the tablet, smartphone, and datacenter growth opportunities as incremental rather than as simply to compensate for a PC decline.
Tablets Should Begin To Ramp, But Not As Quickly As Hoped
According to Intel's CEO, Brian Krzanich, the company has about 50 design wins for Bay Trail-T, with roughly half of them being "two in one" devices. However, while this does sound impressive, and I was particularly heartened to hear that about half of them would be Android tablets, this "announcement" came with the caveat that only about 8-10 models would be on the shelves by Black Friday - the rest would ramp in early 2014.
Now, while this is certainly better than nothing (and much better than what we saw with the initial ramp of last year's "Clover Trail"), what's even more surprising is that it was revealed that "Bay Trail" was largely a pull-in from 2014. While I had actually commented on this back in November 2012 (when it didn't seem clear that Intel would pull it in), this does raise a longer term concern.
While NVIDIA (NVDA), Apple (AAPL), and the rest typically put out one major new chip per year, Qualcomm (QCOM) - Intel's fiercest competitor - seems to be moving at an accelerated clip. Indeed, Qualcomm launched dual core system-on-chip products based on its Krait 200 core (built on TSMC's (TSM) 28nm LP) in early 2012, then a quad core variant by the end of 2012. It then launched an update known as the Krait 300 which found its way into system-on-chip products during early to mid 2013. Finally, Qualcomm yet again updated its processor core with the Krait 400 (a tweaked Krait 300 built on a TSMC's new 28nm HPM high-K metal gate process), which is available in designs such as the latest Amazon (AMZN) Kindle Fire HDX, the LTE-Advanced version of the Galaxy S IV, with more phones/tablets likely on the way.
In the time it took Qualcomm to go from Krait 200 -> Krait 300 -> Krait 400 and the various SoC implementations around them (including an in-house, leadership GPU), Intel has gone from its 5 year old Saltwell in a single core configuration during 2012, to a dual core Saltwell in 1H 2013 to its latest, up-to quad core Silvermont by the very end of 2013. Intel has committed to yearly processor core updates, which should help things along tremendously, but the real question is whether Intel can outrun Qualcomm (Intel's main competitor in the merchant SoC space)?
Right now, it is my belief that "Silvermont" is a superior processor core from a performance/watt perspective to Qualcomm's Krait 400. That being said, Intel will only have "Silvermont" to fight with until likely Q3/Q4 of 2014 at which point it will introduce "Airmont" on its 14 nanometer process. While "Silvermont" can fend off the Krait 400, could it fend off an architecture update on the 20nm process? If Qualcomm is able to get a 20nm Krait-based SoC out in the first half of 2014, then this will likely eliminate Intel's performance/watt advantage on the CPU side of things. If Qualcomm's 20nm Krait hits in Q3 while the 14nm Atom part from Intel hits in Q4, then we'll probably end up seeing a repeat of this year - Intel misses the window for the key designs such as the Amazon Kindle Fire, the Google (GOOG) Nexus 7, and others.
So, can Intel accelerate these efforts? Over time, yes, but in the nearer term? I'm not sure.
14 Nanometer Delayed
It would appear that Mr. Daniel Nenni from SemiWiki was correct in calling for a delay in Intel's 14 nanometer production. While Intel had reassured investors and press alike that 14 nanometer was on track for production as recently as IDF 2013 in San Francisco, apparently there was an issue with defect density. Now, before I get into this properly, I'd like to talk about two different, but related, concepts in semiconductor manufacturing: defect density and line yield:
- Line Yield: the number of good wafers that come out without being scrapped. This is a measure of how effective the material handling, process control, and actual labor.
- Defect Density: This refers to the number of "defects" on the finished semiconductor wafer per given unit of area. The higher this is, the less likely the individual semiconductor dies on the wafer are to be fully functional, and the greater your manufacturing cost per die
CEO Krzanich had this to say with respect to the 14 nanometer delay,
Sure. It's absolutely not the latter. It was simply a defect density issue. This was on the issue -- as we develop these technologies, what you are doing? You are continually improving the defect densities and those resulted in the yield, the number of die per wafer that you get out of the product and what happened as you insert a set of fixes in groups, you will put four or five, maybe sometimes six or seven fixes into a process and group it together, run it through and you will expect an improvement rate occasionally as you go through that. The fixes don't deliver all of the improvements [stock], we had one of those.
Why do I have confidence [that the issues have been overcome] ? Because, we have got back now and added additional fixes, gotten back onto that curve, so we have confidence that the problem is fixed, because we have actually data and defects and so that gives us the confidence that we are to keep moving forward now and that happens sometimes in these development phases like this, so that's why we are going to over it a quarter.
So, it looks to me that the process just ran into a last-minute snag, which caused a quarter's worth of a delay. That being said, I do wonder if this pushes out the 14 nanometer Atom parts, or if this really does just lead to a push out of the Ultrabook oriented Broadwell. We will know more at the investor meeting.
Server Division: Double Digit Growth Is Back...Sort Of
Intel's double-digit datacenter growth story has finally returned - at least on a quarterly year-over-year compare. In the most recent quarter, thanks to the Avoton (microserver part) and Ivy Bridge-EP (2 socket server) parts beginning to ramp, the company saw a healthy return to double-digit year-over-year growth. I was encouraged to hear that cloud revenue was up 40% Y/Y, storage was up 20% Y/Y, and HPC (driven by Xeon E5 and Xeon Phi) were up 27% Y/Y. Heck, even the flailing enterprise server market was up "low single digits" year-over-year. These results are pretty good, but they're not as much as I'd expected.
Indeed, while management was confident that the company would hit low double digit Y/Y revenue growth for the datacenter group, it looks to me that the CFO is hedging slightly and promising to be within "spitting distance" of that number, so my guess is that it's worth expecting 9-10% growth for the year. This is still pretty good, but it's a far cry from the 15% 5 year CAGR given at the 2012 investor meeting. I am hopeful that there will be a continued acceleration of this business beginning in 2014.
That being said, the server business seems to be executing well product wise and revenue wise, so there's not too much more to say here.
Other IA Is In Serious Need Of Revenue Growth
The biggest sore spot this quarter was the "Other IA" segment. This division includes the following businesses:
- Intelligent Systems Group (embedded applications)
- Mobile Communications Group (cellular baseband, RF, and power amplifiers)
- Tablet group
- Phone group
- Service Provider Group
- Netbook group
Management had this to say on the call,
What you see going on in other IA is, first and foremost the embedded or the Intelligent Systems Group had a very good quarter, record revenues, robust growth rate, very profitable, offsetting that is really three things that impact the operating profit. First, netbooks are coming to the end of their life as category. We are down to some single-digit millions of revenue from netbooks. A year ago, that was pretty robust and those were pretty profitable products.
A year ago, if you recall, we were talking about a buying pattern from one particular customer in modem space that didn't repeat this year, so you get a little bit of a year-on-year comp and then the other impact is this is a place where we have been making significant investments.
Going forward (think 2014), I'd expect that the Intelligent Systems group continues to grow, the tablet group grows nicely, and even the smartphone group sees a good uptick. Since the netbook market is now mostly gone, the Y/Y comparisons during 2014 should look a lot better (low base + not much room for further netbook declines). I'd also expect that as the Mobile Communications Group starts to roll out competitive smartphone apps processors and LTE baseband/RF, the revenues there should pick up nicely. Right now, the 3G space is suffering both a suffering TAM and far too many competitors.
But the real question is when this division is going to be profitable. Right now, this division is costing Intel's shareholders north of $2B on an annualized basis thanks to very high R&D spend coupled with a fairly low revenue base. I also would expect that gross margins here are below corporate average. So, assuming that the opex in this space is $3.5B ($2.5B R&D, $1B SG&A), and assuming 50% gross margins, to break even Intel needs to ramp this business to a run rate of about $7B/year, or nearly a doubling of revenues. This is certainly achievable over the next year or two, and once the division is breakeven, there will be a DRAMATIC increase in the company's earnings per share. Understand that this division alone costs Intel shareholders over $0.40/share annualized! It would be even better if this division could turn a profit.
What To Expect From The Share Price
While I don't have a crystal ball, I do believe that this earnings report means that Intel is dead money until its November investor meeting. I don't expect the sell-side to look kindly upon the FQ4 guide, which significantly missed estimates. Further, this guide means that Intel is on track for a 1.3% Y/Y decline in sales. While it was well known that 2013 would be a transition year, it's discouraging that Intel couldn't even stay flat or even eke out a point of growth. While the issue here is end markets moreso than execution, it is nevertheless discouraging.
My guess is that the shares will have support at the 4% yield level (~$22.50), but as always, no real guarantees. In addition, the macro environment serves as an extra liability on top of what will likely be a pack of very frustrated investors. If you believe in the long term story (as I do), then holding is probably the best course of action. But if you're looking at deploying fresh capital, it gets trickier. If you want short-term returns, this isn't going to be the stock for you. If you want to buy the long term story, look for an opportunity once the stock shows signs of technical health. There is no rush at this point, so you can probably take your time searching for the right entry price.
The bottom line is that this is the third time this year that Intel investors (including yours truly) have been kicked in the teeth. While Q3 was in-line on the top and slightly ahead on the gross margin front, the Q4 guide fell flat. While there is room for that Q4 guide to be revised up if Bay Trail system sell-through is particularly good and OEMs are willing to build more aggressively for the designs launching in Q1, I wouldn't bet big money on it at this point. Intel's one-quarter-out guides have been very accurate this year, and I have little reason to believe that I know more than management does.
I am frustrated, but I do believe 2014 will be much better, and I'm going to hold my shares. I still think there's a growth story to be had beginning in 2014 and I'm going to see it through (and, hopefully, for many more years beyond this), but it's going to be another quarter of being "dead money" (with perhaps a rumor or two that Apple is going to dump Intel in the Macs) before the shares get interesting again.
In the meantime, I'll keep you all posted.