Over the last five to six years, microprocessor giant Intel (INTC 1.43%) talked quite a lot about its efforts to become a major player in the market for processors and related components targeted at smartphones and tablets.
Followers of the company's mobile strategy know that it was a spectacular failure. Intel couldn't deliver competitive products at the breakneck pace required to succeed in the market. As a result, as part of restructuring actions announced earlier this year, Intel cancelled all of the mobile processors that it had previously told investors to expect to come to market.
At this point, it looked as though the company had effectively exited the market for smartphone processors altogether, opting instead to focus on building stand-alone modems with the aim of winning business at Apple (AAPL -0.05%).
However, at a recent investor conference, an Intel executive indicated that Intel would eventually be back in the smartphone processor game.
In this column, I would like to outline what I believe would be a winning strategy for the company in the mobile market.
A three-pronged approach
Intel's ultimate goal in mobile is to generate the heftiest return possible for its shareholders; this means keeping spending down and revenues (and gross profit margins) high. To that end, the company can go after this market in the following three key ways: contract chip manufacturing, sales of stand-alone modems, and sales of integrated applications processor and modem products.
Let's take a look at what Intel's strategy could be in each of these segments.
Contract chip manufacturing, but not for everybody
I don't think Intel is particularly well set up to become a truly general purpose contract chip manufacturer. Where I think Intel can add value is for customers that are potentially willing to pay more in order to get better performance/power consumption than they otherwise could have from one of Intel's competitors (assuming that Intel's technology can remain superior to what the competition offers).
Intel already announced that it will build chips for LG, and there are rumblings that Intel may be trying to win at least some of Apple's A-series chip business at some point down the line.
I don't think Intel should (or, frankly, could) build chips for all of the mobile chip companies. However, if Intel is serious about this business, and if its technology is really ahead of the competition's, then going after high end mobile chip customers -- particularly ones that don't compete with Intel's own product lines -- could be a nice way for Intel to profit from mobile.
The stand-alone modem opportunity
When the tear-downs for Apple's upcoming iPhone 7 come in, many expect that Intel will supply modems for some of the models. As far as smartphones go, Apple is really the only company that still relies on stand-alone modems rather than on modems integrated into the same piece of silicon as the applications processor.
As long as Apple continues to use such stand-alone modems, and as long as Intel builds competitive modems for this market, the iPhone could represent a nice opportunity for the chipmaker to try to profit from the mobile market.
That said, this is the most "fragile" of the mobile opportunities in front of Intel. If Apple ever builds up a competent modem team and decides to integrate the modem into its applications processors, then Intel's iPhone-related revenue will go to zero within a couple of product generations.
Integrated applications processors and modems
As Intel's modems gain credibility in the marketplace, there may be an opportunity for the chip-maker to take those modems and integrate them into applications processors targeted at low-end and mid-range smartphones.
In fact, I would imagine such applications processors would be fairly generic, packing ARM Cortex A-series processor cores, graphics and media engines from either ARM or Imagination, and a few in-house technologies (such as the image signal processor, memory controller, and modem).
If these chips can meet the performance/power/feature requirements of low-end and mid-range smartphones, Intel could probably bring in a reasonable amount of revenue for a relatively small incremental investment.
Intel shouldn't bother trying to build in-house chips for the high-end of the smartphone market. These chips tend to be more difficult/complex to produce and the market for them is quite small and increasingly dominated by in-house processor designs. Going after this segment via foundry is the smarter approach.
Why this strategy could succeed
Intel's prior strategy was to try to build products targeted at every segment of the smartphone market while relying on in-house technology. The strategy, for a number of reasons, clearly failed.
What I have outlined above would have a much better chance of succeeding simply because it plays to Intel's strengths. Intel is a world-class chip manufacturer and it is becoming increasingly good at building modems. Attacking the high-end of the market by building other people's chip designs for them solves several problems that Intel faced previously:
- One problem that Intel faced is that it simply couldn't read the market properly and as such didn't wind up designing chips that were competitive in that segment. By letting mobile experts worry about defining the chip's features, Intel no longer has to deal with trying to predict future market trends.
- Another problem that Intel would have faced at the high end of the smartphone market, had it actually built competitive products, is that much of this market is held captive by Apple's own A-series processors, as well as in-house efforts from Samsung, Huawei, and others. Intel isn't going to win over Samsung (Samsung is a direct chip manufacturing competitor to Intel), but it could conceivably win over other customers.
For the low-end of the market, Intel could leverage a lot of third party technologies in order to bring chips to market fairly quickly. Intel's main points of differentiation here would be the modem as well as the performance/power characteristics of its manufacturing technology.
In this case, Intel would again have the benefit of not having to try to predict the future as it would presumably work with ARM and other third-party technology providers to do much of the heavy lifting. Intel's job would be to essentially bundle these technologies with its modems and manufacture them.