After Intel's (INTC +0.48%) analyst day, a number of analysts came away thoroughly unimpressed with Intel's plans in the mobile space. While it really does seem that Intel is setting itself up for long-term success, it is important to understand the other side of the coin -- investing requires an ability to understand both sides and then to pick one.
Indeed, Doug Freedman from RBC Capital Markets downgraded Intel from "buy" to "neutral," noting that it could potentially make more sense to simply take out low-interest debt and buy MediaTek as a way to quickly get into handsets and tablets.
Does this make sense?
Freedman's argument is that Intel would be better served to not spend more than $2.5 billion per year on research and development in a bid to capture the mobile market and to embrace ARM (ARMH +0.00%) by buying a successful player in the field. While this would certainly gain Intel revenue and customers immediately, it's tough to see why Intel would be interested in the company from a technology standpoint. In fact, it makes no sense.
Intel has all of the resources in-house to build competitive systems-on-chips, graphics cores, CPU cores, and all of the other IP that goes into a chip. While the company does need to build up the skill of its design teams and to tweak its design methodologies in order for the company to be better suited for the rapid iteration required in this space, there's nothing that MediaTek has that Intel fundamentally doesn't.
In fact, MediaTek licenses off-the-shelf CPU and GPU IP -- the former from ARM and the latter from Imagination Technologies (IMG +0.00%) and ARM. It assembles it into a proper SoC, and sells it for dirt cheap. The barrier to mimic MediaTek is quite low for a company like Intel. But Intel goes several steps further as it designs its own CPUs, GPUs for some SoCs, and collects its own foundry margin.
Why Intel buying MediaTek would be disastrous for many players
When MediaTek wins, ARM, Imagination, and the major foundries win. While MediaTek's apps processor business is significantly smaller than Qualcomm's, it still commands 11% of the smartphone market -- nothing to sneeze at. If Intel were to outright buy MediaTek, the following would happen within a product cycle or two:
- Intel would immediately replace the ARM cores in the MediaTek SoCs with Intel cores -- as it is doing with its own SoFIA designs. That would rob ARM of -- you guessed it -- 11% of the entire smartphone market
- Intel would either move to its own GPU IP, which it plans to do at the high end with its own smartphone/tablet chips, or it would likely shift all GPU IP to Imagination Tech's from the split between ARM and Imagination Tech. Intel is Imagination's largest shareholder.
- Intel would eventually move MediaTek's chips to its own manufacturing processes, which means that the foundries lose a significant amount of business
Presumably from there, Intel would leverage its manufacturing scale to drive rapid share gains for its newly acquired MediaTek division, which means that all of the aforementioned effects would be amplified to an uncomfortable degree for ARM and the foundries.
Foolish bottom line
Fortunately for ARM, Intel is unlikely to buy MediaTek -- it just doesn't make a whole lot of sense. Analysts are simply impatient or simply don't believe that Intel's product road map is competitive enough (though from what this Fool has seen, Intel's road map becomes very competitive throughout 2014). The Street is rightfully skeptical, although the success of Bay Trail from a technical standpoint should have had people convinced otherwise.
However, the effects mentioned here may simply be delayed even if Intel doesn't buy MediaTek, as they would happen once Intel is a much more forceful player in this space. It will be interesting to see it all play out over the next few years, regardless of the outcome. However, don't hold your breath for Intel to lever up its balance sheet to buy an SoC integrator -- it's just not a great long-term use of money.
