I've been doing quite a lot of work in trying to figure out exactly how Intel (INTC) is going get out of its PC-decline rut, and it seems that Intel can essentially stabilize its client MPU revenues by simply growing extremely quickly in the tablet market. While, of course, we can debate some of the assumptions that I make here, the point is the tablet market is large enough, and the high end of the PC market resilient enough, to allow Intel to actually return to client growth during 2015.
So, How Are You Modeling This?
The PC market today is roughly 300 million units, 82-83% of which belong to Intel, with the remainder belonging to AMD. While it is tempting to assume that Intel's share will eventually be more like 90% given the structural problems (brain drain, compressed R&D budget, focus on too many areas, etc.), let us assume that for now the market share picture stays more-or-less stable.
For the full year of 2013, Intel's PC Client Group registered $33 billion in sales. Assuming that IDC's numbers for 2013 are good (~314 million PCs shipped), and assuming further that Intel had 82% share, this works out to 257 million PC platform units shipped. Now, the PC Client Group revenue number also includes things like discrete Wi-Fi adapters and wired LAN adapters but for the sake of conservatism, we will establish an upper bound on the average PC platform selling price. In this case, 257 million units and $33 billion in revenue suggests an ASP of $128 (for MPU + chipset).
Now, one may naively assume that a tablet cannibalization would lead to the substitution of a $128 platform for a $20 platform, but the $128 number is actually questionable considering the decline in the PC market has actually been at the low end of the market, and has almost certainly come from the decline of notebook PCs, not desktops. So, we have to do some more math and use the data from IDC about the 2013 breakdown between notebook and desktop sales.
In particular, we have the following data:
- Desktop PCs comprised 135.6 million units in 2013 (down from 148.3 million in 2012)
- Notebook PCs comprised 178.6 million units (down from 201.1 million units)
- Intel reports that its desktop platform volumes were down 2% and ASPs were actually up 6%.
- Intel reports that its notebook platform volumes were down 4% as were ASPs.
Now, the IDC data suggests some pretty dramatic market share loss on AMD's part in both desktops and notebooks, but we now do have some idea of how we can model this tablet cannibalization of the notebook (one could argue for notebook cannibalization of the desktop, but those notebooks are unlikely to be cannibalized by the tablets).
The unfortunate part here is that we actually don't have enough information to do a mathematically correct analysis, but for the sake of simplicity, let's simply assume that the $1 billion lost was due to a 3% decline in Intel's PC units, which went from ~279Mu (assuming 80% market share) to ~257Mu (~82% market share). This means that Intel lost $1 billion in sales on 22 million units, implying an average selling price of the cannibalized units of just $45/unit on a platform level.
This Changes Everything
If you are assuming that Intel loses a $128/unit sale to a $20 processor, then you'd be correct in assessing that Intel is in deep trouble. But what we're really looking at is a $10-$20 single-chip tablet platform 'cannibalizing' a $45-$50 platform which - until very recently - was a clunky two-chip solution based on large, disabled die.
So, let's assume now that Intel is replacing a $50 platform with a $15 tablet. While Intel still needs to sell 3x the units to "make up" for that sale, Intel's higher margin PC business (i.e. Core i3 or better) is still safe from this cannibalization. The cannibalization has happened at the low end and the volume left to cannibalize is probably far less than you'd think. In fact, even IDC believes that PC shipments will stay above 300 million units by 2018.
Assuming that Intel loses another, perhaps, 20 million units @ $50 a piece, Intel needs to only sell 66 million tablet chips at $15/piece. I suspect that as Intel brings down the bill-of-materials for SoFIA and Broxton and as it integrates more stuff (to drive added value), Intel can probably stay at the $15 ASP level but grow units to north of 100 million units. At that point, even a very pessimistic forecast for the PC can be blocked by tablets.
This would also ultimately mean that the highly profitable server group as well as the incremental opportunity in smartphone apps processors/phones will drive real growth on top of the current revenue base. Intel will see legitimate growth come 2015 as the contra-revenue disappears on the tablet side of things, smartphones begin to ramp, and server/cloud growth marches on.
Some Numbers & A Price Target
Of Intel's $52.7 billion in sales seen during 2013, $33 billion came from PCs. Assuming that PCs + tablets stay flat at that $33 billion level during 2015, and assuming 5% of the smartphone apps processor market (~70 million units+ @ $10 ASP) and ~15-20% of the cellular baseband market (up from 12% in 2013 as XMM 7160/7260 ramp), all with 12% growth in the data-center group in 2014 and 2015 (I do not believe Intel's 15% CAGR), Intel could be doing north of $56 billion in sales in 2015.
I also expect opex to be flat/down from 2013 levels, and I expect gross margins to tick up to 62-63% as Intel's tablet/phone efforts will drive greater utilization. With a 27% tax rate, this amounts to $11.75B in net income or $2.36/share assuming 4.98 billion shares outstanding.
Applying a 14x multiple to this would suggest that Intel could be trading north of $33 if Intel gives guidance in Nov. 2014 roughly in line with the assumptions given here. This would suggest upside of 35% from current levels over the next year or so. I also did not assume material impact from the growth of software & services, nor do I factor in growth in NAND (but the DCG growth captures enterprise SSD sales growth). I also expect no material foundry revenue during either 2014 or 2015.
After running the numbers here, it's clear to me that as long as Intel can execute, and if my math on the low-end PC unit ASPs is even remotely accurate, the stock could very handsomely reward investors over the next year. However, execution risk is quite real, and Intel may choose to guide conservatively during the Nov. 2014 meeting with the intent of "surprising" during 2015, so keep this in mind before you (particularly the options fiends) place any bets that depend on precisely timing a potential move.