Investors and analysts following Intel's (NASDAQ:INTC) mobile efforts are probably aware of the incredible amounts of money that the company is spending in a bid to become a major mobile player. During 2013, Intel reported that its Mobile and Communications Group ("MCG") lost nearly $3.2 billion, and the loss is only expected to get wider during 2014 as the company provides "contra-revenue" support for its tablet platforms.
However, despite these significant losses, Intel still plans to stay in this market. At an investor conference last week, Intel's CEO Brian Krzanich laid out some reasons its mobile investments aren't merely an incremental opportunity, but in some ways required in order to defend its share in some parts of the PC market longer-term.
Connectivity is the name of the game, even in PCs
The vast majority of laptop computers and a great deal of the tablets sold today do not contain cellular connectivity. However, Krzanich made an interesting remark at the recent Citi Technology Conference (emphasis mine):
If you look at the computing ecosystem, and we think that most of the devices that you have today are going to need a spectrum of comms, whether it's Bluetooth or Wi-Fi or 3G or LTE, thinks like Bingbooks and Chromebooks are going to be more and more connected. In fact, they're really only functional -- or highly functional -- when they're connected. We think that 25 to 50 percent of those over the next few years will have 3G or LTE. So even in our core business, you've got to have that comms and mobility capability.
Now, realistically, for products like Chromebooks, Intel could very easily use an off-the-shelf, stand-alone cellular solution from competitor Qualcomm. In fact, even though Intel develops its own Wi-Fi solutions for PCs, oftentimes, Intel-equipped PCs will ship with solutions from third-party vendors (for example, Apple's MacBook Pro laptops use Broadcom Wi-Fi chips).
So, if Intel could just use a third-party solution, why does it care so much about building all of these connectivity solutions in-house?
Integration and scale required
One of the long-term trends in the computing industry has been toward integration, particularly as end-markets mature. Reinforcing that point is the following slide shown during Broadcom's presentation at the Citi Technology Conference:
The idea here is that early on in a particular market (for example, PCs), there are a lot of different chips dedicated to carrying out a particular function. Over time, however, as the market evolves and then reaches maturity (as the PC market arguably has), most of that functionality gets integrated into a single chip known as a "system-on-a-chip."
For example, look at a modern PC processor from Intel today. It integrates a graphics processor, CPU cores, a memory controller, and an L3 cache. Each of these components was, at one point in time, found on a separate chip.
So, as connectivity in the computing market becomes more essential, it only makes sense that more of that technology will be absorbed into the main processor.
Further, while this integration is often done for legitimate technical reasons, there are also economic factors to consider. To illustrate, direct your attention to the following slide from processor IP vendor ARM Holdings (NASDAQ: ARMH), which shows the kind of cost savings that can come from integration:
Note the key point ARM makes: Integrating the functionality reduces the cost for the customer buying the final chip and improves profitability for the chip vendor.
It's not hard to see how integrating cellular and connectivity functionality into a PC chip targeted at Chromebooks makes sense. Chromebooks are inexpensive (which means that the lower the component costs the better) and, as Krzanich pointed out, can't be used to their fullest without Internet access.
If Chromebooks really catch on (and CNET reports the folks at Gartner think that Chromebook sales will hit 14.2 million units by 2017), then Intel could be at a pretty significant disadvantage if it can't offer a fully integrated solution like its competitors can.
Wait a minute -- weren't we talking about mobile?
Circling back, it's worth tying this discussion in to Intel's need to play in the mobile market.
The plain and simple reason Intel needs to play in mobile is that this market drives a lot of volume. To quote Krzanich at the same Citi conference:
I mean just the economics of shrinking the die and you have to have enough die to produce to keep our factory full. And if you're shrinking them in half every two years, you've got to have enough die -- enough growth -- to keep your factories full. So you need scale. I think scale is the number one thing that you have to have.
In all likelihood, Intel probably isn't going to see tremendous unit volume growth from the PC market longer-term; even multiple indications suggest the market is stabilizing. The datacenter, while a very nice revenue and profit growth engine for Intel, also isn't likely to see the kinds of huge volumes needed to keep the company's future factories full.
On the other hand, according to IDC, 1.2 billion smartphones will be shipped this year, and as many as 1.8 billion will ship in 2018.
Further, these smartphone chips are pretty reasonably sized. For example, Apple revealed that its just-launched A8 chip weighs in at 89 square millimeters, and Intel states that just-launched Core M processor comes in at 82 square millimeters.
So, if Intel needs to develop much of the processor, graphics, and communications technology for low-cost PCs anyway (Intel's tablet-oriented chips and its low-cost PC chips already share a tremendous amount of technology, for example), it makes sense for Intel to keep trying in mobile until its products in this space are competitive enough to profitably gain share.