Over the last couple of years, I've come to realize that stock prices are almost entirely predicated on future expectations. That's why the traditional metric of "trailing P/E" that people use to try to determine whether a stocks is under/overvalued is a little bit misleading. A P/E of 100 isn't terrible if earnings are expected to double each year for the next 5 years, but of course shareholders get burned if these expectations don't materialize. In the case of Intel (NASDAQ:INTC), I believe the exact opposite is true - analyst estimates for 2014 are too low, and as a result I see major upside ahead over the next 52-weeks.
Analyst Estimates: What Are They, And Why Are They Too Low?
According to Yahoo Finance, here are the estimates for the end of fiscal 2014:
The mean prediction is a 4.3% Y/Y revenue increase from fiscal 2013. There are two problems here:
- The mean estimate for 2013 is too low at $53.58 - this is significantly below the "low single digit" revenue growth guidance that was given in the full year guidance, so the base from which to measure the growth is off.
- The 4.3% projected growth from 2013 is more rational, but still a fair bit off from what I believe is much more likely to be a high single digit/low double digit growth year.
In addition to the 2013 estimates coming in well below the actual guide (note that this presents an opportunity for upside surprise), the 2014 estimates seem to fail to consider the numerous growth drivers that will be in play during the full year of 2014, which I will outline here.
2014: A Full Year Of Leadership In Tablets
At the end of 2012, Intel released its "Clover Trail" SoC, which was basically two of the old 32nm "Atom" processor cores stitched together in a new, low power SoC for the tablet market. While early sales are very encouraging, with sales numbers suggesting that Intel now has roughly 6.75% of the branded tablet market, this product has yet to penetrate the Android tablet space in any major way. I believe that with the release of "Baytrail" based on the "Silvermont" core, Intel not only secures its leadership in Windows 8 tablets (which should gain share in the broader tablet market), but it also makes a significant incursion into the Google (NASDAQ:GOOG) Android tablet market.
In 2013, "Baytrail" will only contribute to Q3 and Q4 sales, while in 2014, "Baytrail" ("Silvermont" SoC) will contribute to full-year revenues, along with its "Airmont" based successor "Cherry Trail".
To try to quantify this, let's assume that the broader tablet market grows at a conservative 50% CAGR from 2012's $2.7B (according to Strategy Analytics), implying CY2014 revenue base of $6B. It seems that, again according to Strategy Analytics, Intel now has ~6.75% of branded tablet market (90% of the Windows tablet market which itself has 7.5% of the total branded tablet market). This was achieved in a mere quarter as "Clover Trail" launched in late Q4 2012.
So let's assume that the Windows tablet market doubles its market share to 15% of the total tablet market, and then let's further assume that Intel has, once again, roughly 90% of this market. Assuming $20-$40/chip (depending on core count, frequency, etc., so let's say ASP of $30), which is what the "Clover Trail" sells for, I could see Windows 8/Blue tablet revenues for Intel of $810M in 2014 just from Windows 8 tablets.
Now, assuming that Intel gets into the Android market in a big way, which comprised 43% of the branded tablet market in Q1'2013 (and likely to continue to take share from Apple (NASDAQ:AAPL)), and assuming Android has ~50% share in 2014 (since the trend is towards cheap, but capable smaller tablets), I could see Intel's "Bay Trail" and "Cherry Trail" parts taking a conservative 10% of this market during the year. This implies an additional $300M in tablet-related revenues.
In 2014, I expect Intel to do about $1.1B in tablet sales.
2014: The Year Of The Intel Smartphone?
Intel has had a smartphone solution out in the market since 2012, but it hasn't really gained traction - Intel still has but 0.2% of the smartphone market. This is actually pretty expected - Intel doesn't have an LTE solution available yet, so all of the merchant chip business goes to Qualcomm (NASDAQ:QCOM) at the high end, while the vertically integrated players like Samsung (OTC:SSNLF) and Apple use (mostly) in-house chips (Samsung actually uses Qualcomm's Snapdragon 600 in the US version of the Galaxy S IV).
Further, the Atom Z2580 (dual core 32nm Atom) is a fast machine, but it's not fast enough to really make smartphone vendors want to switch from their trusty Qualcomm, whom they would need to go to for the modem anyway. But in 2014, things change rather dramatically.
Intel Gets A Performance Per Watt Advantage
While in 2012/2013, Intel's dual core/quad thread Z2580 was mostly on-par with the quad core Qualcomm chips in terms of CPU performance thanks to much better single threaded performance and hyperthreading, it didn't really offer that much of a performance per watt advantage...if any. But if Intel's presentation on "Silvermont", the ground-up redesign of Atom for phones, is to be believed, then a dual core "Silvermont" will have a massive performance/efficiency advantage over the quad core competition in SpecInt 2000 (which is a benchmark that ARM (NASDAQ:ARMH) uses to compare its cores, so no foul play here):
Vendors will be more than happy to sell smartphones with better battery life and better performance, and if Intel really delivers with much lower power and better performance, then this is a very clear "win". But the apps processor is only half of the equation...
Intel Gets LTE
As I said, something that has really held Intel (and everyone not named Qualcomm) back in smartphones is that Intel lacked an LTE solution. Well, luckily, not only is Intel shipping a multimode solution to customers by the end of the first half of 2013, but according to Signal Research Group, Intel's LTE solution was the top performer in LTE, implying that the upcoming LTE modem from Intel isn't just a "me too" LTE product, but a real winner. This only improves Intel's chances of being able to sell the complete platform to smartphone vendors.
LTE And "Silvermont" Phone Chip Coming In Q1 2014
The best part about this is, just as tablets will contribute to revenues for the full year of 2014, so will smartphones. According to a presentation at IDF 2013 Beijing, Intel's "Merrifield" (the name of the smartphone chip) will be on the market in Q1 2014:
So, get this. Intel will be selling a ~$20 apps processor, while at the same time selling another $10 - $15 modem. Now, you will see critics saying that "oh, the market is dominated by Apple and Samsung so Intel doesn't have a shot", but this is completely unfounded. Qualcomm took 43% of the smartphone applications processor market in 2012, and it's not in the iPhone, and it's only in some editions of the Galaxy S. Further, the market as a whole in 2012 was a $12.9B one.
Let's then say that Intel gets, perhaps, 5% of this market in 2014 thanks to "Merrifield" and the integrated apps processor and modem chip called "6331". Assuming that the market grows another 50% in 2013 and then yet another 40% in 2014, we are looking at a $27B market. 5% of that implies $1.35B in incremental revenue for the year, with further share gains more than likely in the years ahead.
Note that these market share shifts can happen very quickly with the right product(s). How quickly do you ask? Well, here's what the market looked like in 1H 2010:
Texas Instruments (NASDAQ:TXN), ST-Ericsson, Freescale (NYSE:FSL), and Marvell (NASDAQ:MRVL) are now nowhere near the top 5, Broadcom (NASDAQ:BRCM) is now in the top 5, and Qualcomm took even more share. The smartphone design cycle is short, and with a really good product from a performance/watt/$ standpoint, share can be taken quite quickly.
I think that Intel may even be able to snatch 10% of the market by FY2014e, but I don't want to be too aggressive.
Further, did you know that the baseband/modem market is even larger than the apps processor market (although slower growing)? Here, Intel actually is in the top 3 vendors (with 12.3% revenue share) and this is with no LTE solution yet shipping in the market!
Assume that this market grows at a 15% CAGR through 2014, and assume that thanks to a solid LTE solution as well as apps processor, Intel grabs 20% market share by then. This would imply a $23.5B market, with baseband revenue of $4.7B by the end of 2014.
Server Market Growth
Intel expects the data-center revenues to get back to a 10%+ CAGR. This is thanks to a fundamental broadening of the chips and other products that Intel sells into this market. Growth slowed in 2012 because the enterprise IT market slowed, but this was offset by growing cloud revenues (40%+ growth there). With the new "Atom" offerings further widening the range of products offered, and with the "Xeon" products moving to lower power, higher performance, and further scalability, I expect market share gains, coupled with a secular market growth story, to accelerate revenue growth in this segment. I'm expecting ~10% growth this year in data-center, with 15% in 2014 as the full impact of the new Atom and "Ivy Bridge-EP/EX" products is felt.
With 10% growth in 2013, and 15% growth from there in 2014, we are looking at FY2014e revenues in this segment at $13.55B, with 50% operating margin, to boot!
PC Client Market - Haswell, Broadwell Bring Us Back To Growth
The PC client market underwent what I believe is a "reset" in 2012. Form factors were no longer desirable in light of the move towards thinner, lighter, longer battery life, and touch. People still very much want performance, but the state of mainstream laptops in 2012 and early 2013 is nothing short of a disaster, although models such as the Lenovo Yoga give us hope.
My bet is that with "Haswell" enabling "all day battery life", thanks to massive improvements in idle power and power management, the high end tablet market goes essentially the way of the dodo as 2-for-1 "Haswell" based convertibles have a significant value proposition. This only gets better in 2014 with "Broadwell", which should bring power even lower and performance in both CPU and graphics higher.
Further, for even lower cost designs, Intel will be shipping a version of its tablet-oriented "Bay Trail" in 2013 and probably a version of "Cherry Trail" in 2014 for the notebook/convertible market. While these will not be as high performance as the "Haswell" and "Broadwell" products, they will still be much faster than the ARM based products (at the same cost most likely), while servicing both the Microsoft (NASDAQ:MSFT) Windows and Google Android markets. But these won't be counted in "PC Client Group" revenues and instead will be in the "Other IA" group.
For the PC client group (which includes convertibles/hybrids with "Core" products on up to monster desktops), I expect 2013 to be roughly 2% up from 2012, and 2014 to be up perhaps 5-7% (as the Haswell/Broadwell impact will be in full effect), since this excludes the Atom which could drive significantly more volume, so that means in 2014 I'm looking for $34.5B in 2013, and $36.2B in 2014 at the low end of the range.
As far as the "Atom" convertibles go, it's hard to really come to a reasonable estimate here, but my best estimate would be to go back to the 2009 sales levels (1 year after the Netbook launched and drove volume) to get a feel for the Atom based convertible sales in 2014, which was about at the $1.5B mark.
Software And Services:
Intel has a bunch of software and services revenue, most of it from McAfee. It's tough to get a grip on where this is going revenue wise, but in 2012 it did $2.4B, up from $1.9B. Let's assume a 15% CAGR exiting 2014, which implies a $3.2B business exiting the year.
This is NAND flash and a few other things, and worth about $1.6B in sales both in 2011 and 2012. While NAND continues to grow, I think growth at the top line will be slower than peers in the space as Intel only wants the higher margin business. Let's be conservative and assume a 7% CAGR through 2014. This gets us to $1.83B in FY2014 sales.
Summing It All Up
I've shown you the parts, now here's the total estimate (FY2014e):
- PC Client Group: $36.22B
- Other IA: $1.5B (convertible/low end PC Atom products) + $4.7B (baseband) + $1.35B (phone apps processors) + $1.1B (tablets) = $8.65B
- Datacenter: $13.55B
- Software and Services: $3.2B
- All Other: $1.83B
- TOTAL: $63.45B
Assuming that Intel isn't taking excess capacity charges, we can expect that gross margins pop right back to 63%, which is where they will be according to management in 2H 2013. Assuming, then, operating margin in-line with that in 2011, this implies operating income of $21.57B and net income of $16B. At a diluted share count of 5B, we are talking GAAP EPS of $3.2/share. At current multiple of 12x EPS, we could see the shares worth $38.4! With a credible mobile growth story in place, a multiple of 13x or even 14x earnings wouldn't be out of the realm of possibility, implying a share price well north of $40 (although expecting dramatic multiple expansion may not be wise yet).
In any case, I believe shares remain dramatically undervalued compared to its long-term earnings power and prospects.