One of Intel's (INTC) biggest pushes for revenue growth is trying to win over the tablet space. During the most recent conference call, CEO Mr. Brian Krzanich reiterated Intel's goal of putting 40M-plus tablets in the hands of consumers by the end of 2014.
I would like to quickly take a glance at the tablet market to see how realistic this goal is and what kind of impact this could have on financials.
A Quick Look At The Tablet Market
In December of this year, IDC released market share data and the 2014 forecast. In the same report the overall tablet market was predicted to be ~228M units, and a projection of ~270M units in 2014.
(click to enlarge)
A separate IDC study shows tablet market share broken down by vendor (data is from Q3).
(click to enlarge)
There is the "Others" category, and in this category is where we see "disposable tablets." There is also another good read on SemiWiki, although it's describing low cost smart phones, which details how manufacturers are able to put these devices in the hands of consumers.
To paint a more favorable picture for Intel based on IDC data, let's assume that Apple's market share falls to 25% in 2014, meaning more sockets are available for Intel. Also, the "Others" category includes companies like Sony that sell Intel powered tablets, so I will assume that ~1/3 of the "Others" category are available sockets to Intel, and that the "Others" market share also falls to 30% (again, more sockets available for Intel).
This would leave around ~150M or so tablets open to Intel next year, meaning that if Intel succeeds in its goal, the company would need to capture over 25% of the total tablet market.
Chipworks pegs the die size for Qualcomm's (QCOM) S800 processor at 118 mm^2, highlighting Intel does have a slight advantage over Qualcomm's high end. However, in the same link, Chipworks highlights the RockChip RK3188. This is a quad core CPU + GPU in ~25mm^2. So while Intel's Bay Trail may have higher performance, much of the competition actually uses smaller die, and sometimes by an extreme amount as we see in the RK3188.
(click to enlarge)
These complete devices are also very cheap. The mini-pc above comes with WiFi, blue-tooth, HDMI out, and the quad-core Rockchip RK3188 SoC for about $80 total, preloaded with Android.
Lastly, Samsung runs its own fab facilities and has its own line of chips. Apple and Samsung are the two largest vendors, with a combined market share of ~50%. While Samsung often uses competitors chips in its products, I doubt the company will be shifting its products to rely solely on QCOM or INTC chips. If the rumors of Apple moving production ever bear fruit, Samsung will likely try and use its own processors more to boost fab utilization.
How Much Cash Would an Extra 40M Units Bring In?
The appeal of tablets isn't as much in the raw processing power as it is in the form factor, weight, display, battery life, etc. Processing power is important to be sure, but only if you do not want your tablet to feel "sluggish." There is probably a small cross section of consumers that would like to have a Core powered tablet, but IDC data has noted that the tablet ASP has declined and is forecasted to continue to do so.
Tablet processors are typically very cheap, but in this assumption I will assume an ASP of $35, which is just shy of the full price of the Z3770 according to Intel's ARK website.
If Intel achieves 40M sales next year, this will generate an extra $1.4B in revenues for the full year, or an average of $350M per quarter. Given that Intel typically generates around $13B or so each quarter, this is about 3% revenue growth.
To exaggerate this somewhat to show why I disagree, let's say that Intel captures all 140M sockets I estimated that the company could truly compete to win. This would translate to just over $5B in revenues for the full year, or 10% revenue growth YoY.
And keep in mind that my estimation above is revenue, not profit.
A Comparison To Intel's Big Core Chips
(click to enlarge)
Benchmarks comparing a 'Y' series (very low power) Core i5 chip against a quad-core Bay Trail tablet chip reveal an interesting data point. The bigger i5 chip has better single threaded performance than the smaller tablet chip, but when looking at multi-threaded benchmarks the gap becomes extremely narrow. This is due to Bay Trail-T being quad core processor.
The biggest difference between the two is likely going to be graphics performance, but the issue here is that the iGPU performance of the low power Core i5 chip likely isn't going to open many doors not accessible by Bay Trail-T. The graphics performance of this Core i5 chip is going to likely be in this "no man's land" of sorts, where it's not quite good enough for newer titles, but is a good fit for very casual gamers. But this category of consumer would likely be just as satisfied by Bay Trail.
Intel bulls are betting on growth, and a large chunk of this growth is centered around Intel being able to make in-roads into mobile.
As a geek, it's easy to get zeroed in on the numbers and metrics that will tell the story of the chip's competitiveness. However, no matter how good a chip is, that's only part of the story. Chapters 2 and 3 are going to be the total market size and how much this could impact financials, because that's what we care about at the end of the day.
Intel is a solid company with a healthy moat, but when you're as big as Intel is, growth doesn't come easy.
If you look at the tablet market as a whole, and make some reasonable assumptions (like factoring out Apple's market share to judge how many sockets are available to Intel), you'll see that even the most optimistic scenarios are likely to attribute more than a few percentage points of revenue growth.
But while Intel is growing revenues via tablets, the company is also making the smaller, cheaper chips that go into these tablets more powerful, so the company runs the risk of cannibalizing the more expensive Core line as consumers start opting for non-traditional form factors.
In my opinion, for meaningful growth, Intel would need to succeed in both smartphones and tablets, as smartphones represent a much larger market. But then will Intel have to implement a plan for smartphones, similar to the "contra revenue" plan for tablets? And if so, what is the time frame for real growth then if these subsidy plans are eating into margins?
I personally would rather see Intel compete more seriously as a foundry. Making a few dollars from serving as a foundry for other chip makers could attribute to meaningful growth while simultaneously taking lunch away from competitors. Intel has stated the company will focus more on transistor density going forward.
If Intel is serious about serving as a foundry, the above graph is showing that Intel plans on overtaking the competition from a density standpoint. If this is the case, Intel plans on being quite a bit ahead in terms of transistor density beginning next year with 14nm products. If the company is successful, that extra 35% density above the competition would help Intel extract more cash from foundry services than the competition.
But in the nearer term (prior to Intel making meaningful inroads as a foundry), I am less excited about growth out of Intel until I see the company competing, and winning, in both smartphones and tablets, and without the assistance of subsidies.