Intel's (NASDAQ:INTC) Q2 results were largely in-line with expectations. Guidance was cut slightly - revenue is now expected to be flat (vs. prior guidance of up low single digits) and GM for the year is expected to be 59% (vs prior guidance of 60%).
The stock's down a bit over 4% in after-hours trading as I write this. I think that's an overreaction - the market's missing the forest for the trees. If the dip persists over the next few days, it's a pretty nice entry point. I'm not going to go over the thesis for Intel here; I established it pretty thoroughly in my previous write-up, which was quite fortuitously published when Intel was trading under $20. Even counting the after-hours decline, the stock's up over 15% since then as the market realized the bottom wasn't falling out. (It was up over 25% in June.)
As usual, the key takeaways from this quarter weren't in the results, but rather in the commentary on the conference call (see transcript for interesting tidbits like Stacy Smith revealing that core/maintenance capex is $7-$8B/yr). I've highlighted a few key points below.
Commentators are already bemoaning that Haswell is a "failure" because it didn't stem the decline in PC shipments. Yet despite that well-publicized drop, PCCG revenue was up sequentially and down significantly less y/y (7.5%) than the 11-and-change percent that Gartner/IDC reported. When questioned on the discrepancy, Intel management noted that the numbers from Gartner didn't count devices like Microsoft's Surface Pro as PCs, yet since such devices still utilize Core-branded chips from Intel, sales of those devices add to PCCG's top/bottom line. To me, this suggests that headline numbers from the research firms should be taken in a grain of salt, especially because Haswell isn't a failure.
I built a desktop with an i7-4770k (the highest-end mainstream Haswell chip) and have been blown away by how cool it runs even when I overclock it and put it under load. But never mind that, because I'm not the typical user, and Haswell wasn't really designed for me. It was designed to increase functionality of ultramobile devices by enabling significantly extended battery life while offering the same (or higher) levels of performance as previous-gen Ivy Bridge chips. While many doubted Intel's claims of major perf/watt gains pre-launch, many reviews of Haswell-powered laptops like the new MacBook Air substantiate the massive gains in battery life. Some users even go so far as to say that the battery life/functionality combo is so exceptional that they're no longer using their iPad:
I used to tote around my iPad for the sole purpose of watching movies/videos on flights so that I don't wreck the battery life on my laptop. I really don't have to worry about that so much now. Even burning through video on a 4+ hour flight, I still have more than enough battery for an evening of catching up on work. I find myself leaving the iPad behind now. Why bother with the added weight in my bag?
Even if this is a purely anecdotal example, Intel executives said something else on the call that's very important: fanless Haswell. They said it twice, actually. For the uninitiated, one of the reasons you can't have a high-powered chip in a mainstream tablet (besides battery draw) is that such chips generate fair amounts of heat, meaning they need to have a fan to cool them. There's no room for a fan in tablets, and even if there were, consumers likely wouldn't be happy with a tablet that got hot and started buzzing like old laptops have a propensity to do. Well, fanless Haswell means that you're bringing the power of Intel's Core lineup to high-end tablets; it doesn't take an expert on microarchitecture to know that Core will absolutely smoke any ARM-based chip that tries to compete with it.
But it gets even better early next year, because Broadwell is going into production in Q4 and launching in the first half of next year. Broadwell is the next tick - the shrink of Haswell to the 14nm process - and it will basically be everything that users already like about Haswell, except better. If some SKUs of Haswell will be fanless, then even more Broadwell SKUs will be fanless, meaning that tablet OEMs can gain a serious advantage on their flagship designs by opting for the higher-performance Intel Core chips.
Intel's Atom is currently viewed as the underdog of the low-power chip world, just like Hugh Jackman's fighting robot Atom was the underdog in Real Steel. You know what happened in Real Steel? Atom kicked serious butt. Intel's Atom will do the same. Management reaffirmed that Silvermont will offer 3x the performance at the same power, or consume 5x less power at equal performance, which would put it ahead of all existing (and slated) designs from the ARMy. Plus, Intel finally has a multimode 4G/LTE solution (better late than never). Atom will enable Intel to compete in the low end of the tablet market; even if it doesn't blow them out of the water, it'll be competitive enough to get Intel some design wins, which they can build off.
And build they will. New CEO Brian Krzanich has made it extremely clear that "ultramobile" (like mobile, except buzzier and hipper, it's what all the kids are saying these days) is Intel's priority. So much so, in fact, that he's explicitly stated that Atom will be brought on the same timeframe as Core, and may even be moved to new nodes ahead of Core in the future.
This brings up an interesting point. Intel's opex has been soaring on account of massively increased R&D spend, most of which is going to Atom. It shows: take a look at Intel's different business units, and Atom has been a pretty significant drag on EPS. If operating leverage on increased unit volume just led to breakeven, it would result in significantly higher income; actual profit would force analysts to completely remodel Intel.
Management was questioned on the historically high level of opex as a percentage of revenue. Their response: the trend will begin to reverse as unit volumes ramp with the roll-out of Bay Trail, Merrifield, and Avoton. This is a double-whammy, because higher utilization leads to higher gross margin as well; CFO Stacy Smith implied that it could head up a couple hundred basis points, and there's still room to run. Intel's 56% GM in Q1 was near the trough of Smith's long-term guidance of 55-65% GM, which is heavily utilization-dependent. As the fabs fill up with new unit volume (and perhaps foundry clients, though that's more long-term), GM should be higher than what we've seen over the past few quarters.
Beyond this, opex will be further reduced by the synchronization of Atom and Core. Per Krzanich:
What you're going to see and remember, as we move our architecture more and more toward SOC, a lot of the work we do bridges between Core and Atom and as we pull Atom forward onto the leading-edge technology, again, that allows a lot of the architectural work to be in alignment and be shared amongst the two products. So, we believe that we can move faster on Atom and into this ultra-mobility space without having to change or increase our OpEx.
Revenues here were up 6% sequentially, and flat y/y. Nonetheless, management reaffirmed their guidance for double-digit growth for the year, suggesting that we'll see a steep ramp in DCG revenues. Management suggested this is partially due to a macroeconomic pickup in 2H, but more related to the upcoming refresh. Management's continued confidence in light of their overall guidance cut suggests that DCG will continue to be a bright spot for the company.
The changes implemented by Krzanich seem to be mostly positive. I would certainly prefer if they hadn't cut guidance, but focusing too much on the next few quarters is missing the forest for the trees. For the first time ever, Intel will have truly market-leading offerings in the ultramobile space; their two-node process advantage over the rest of the industry will only continue to widen as competitors pause to implement FinFETs. Intel is poised to take significant market share in tablets and other ultramobile devices. 2014 should be a major year for Intel; from an investor's perspective, changing market sentiment and recognition of Intel's mobile momentum will arguably be as (or more) important than the fundamental improvement in the financials.
Ultimately, though, a nice story only matters if it's complemented by a nice valuation. Building a discounted cash flow model factoring in management's guidance for reduced capex as well as moderate revenue growth based on share gains in tablets/ultramobile, decreasing opex as a percent of revenue, and increased GM on account of increased capacity utilization, I come to a fair value of $27 for Intel shares today (though with more aggressive assumptions on revenue growth, gross margin, and opex/capex reductions, you can push it up to $30 give or take). Even the conservative valuation represents over 16% upside from today's after-hours price. Keep in mind that this represents fair value today; a one-year price target would be higher ($28-$29 sounds about right). Also keep in mind that this is a somewhat conservative valuation that does not factor in incremental opportunities like Intel Custom Foundry, Intel Media, etc; it also discounts cash on the balance sheet by 25%* and assigns no value to Intel's equity investment in ASML. (Combined, those two adjustments knock off about $1/share in fair value; fully counting the cash and counting the ASML investment at 50% of current market value, you can make the case for INTC shares being worth $28 today. I, however, prefer to be conservative.)
All told, even simpler earnings projections and EV/EBITDA comps suggest that Intel is pretty undervalued. The market's flawed perception of Intel's market environment and apparent disregard for its technological lead offer a nice entry point today (with the caveat that it's not quite as nice as Intel at $19-$20).
*You may be wondering why I choose to discount cash on the balance sheet. I realize this is not standard practice, but I do so because I don't view a dollar of cash on the balance sheet as a dollar of cash in my pocket. Why? Well, partially to account for repatriation, but beyond that, there are three primary ways corporations use excess cash: a) pay dividends, b) buy back shares, c) make acquisitions. In scenario A, if the company pays out a $1 dividend, the company is worth $1 less, but I'm only worth $0.70 or so more, since I have to pay taxes on that dividend. I just lost 30% of the supposed "value" of that cash; discounting it helps defray that. With regards to scenario B - companies are historically poor at timing share repurchases. With regards to scenario C - companies have a tendency to overpay for acquisitions. All told, since I'm not guaranteed a dollar in my pocket for every dollar in the company's account, I choose to discount it. I figure there's no harm in being conservative - if the company looks significantly undervalued even after a conservative valuation, and true value turns out to be higher, then returns will be greater than expected. Better to underpromise and overdeliver than the other way around.
Disclosure: I am long INTC.
Additional disclosure: The views expressed here are solely my own and should not be taken to represent the views of my employer. Seeking Alpha never holds a "house opinion."