I'm not an expert in semiconductors or Intel (INTC) and know little aside from what I've learned from authors on SA. However, I do think sometimes there can be value in looking at things from a fresh and non-technical perspective. Intel's recent decision to open up its foundry as well as bribe - sorry, assist - tablet manufacturers in redesigning their devices so they can use Intel chips, is worth commenting on from the point of view of Intel's business strategy and what it tells us about the competitive situation that Intel now finds itself in.
I found the reaction of the market to Intel's latest investors' Day rather ironic, as Intel's stock rallied when Intel announced it would be opening up its foundry and then lost ground when flat revenue projections for 2014 were given. No one wants to see flat revenue, but it would seem to me that the news about foundry and contra-revenue for tablets was of far more importance, and not in a good way.
Why is Intel subsidizing its mobile efforts now rather than a couple of years ago? Did Intel believe that its chips from the last generation - Clover Trail - were so good that they could win on their own merits? What does this say about Intel management's ability to read the competition and understand where their industry is heading?
Intel's message to investors over the last few years has been: trust us; we can win in mobile, we have the best design teams and the best process technology. In the light of their latest decision, it will be hard now to take anything Intel says seriously.
Why has Intel struggled in mobile?
There aren't any inherent advantages when it comes to power consumption for Arm (ARMH) chips, which use the RISC architecture, compared to Intel's x86 chips, which use the CISC architecture, so there's no technical reason why Intel chips shouldn't be able to compete.
There must be other reasons, and the contra-revenues, aka the tablet chip giveaway bonanza, is illuminating in that respect:
Tablet makers have designed their devices around chips, (Arm chips as it happens, but the instruction set architecture isn't the issue) from the likes of Nvidia (NVDA), Qualcomm (QCOM) and Samsung (GM:SSNLF), and these chips have been specifically designed with mobile connectivity in mind so they already have much of the connectivity functions you'd expect a tablet to have, such as Bluetooth and Wi-Fi, integrated into the chip itself.
Intel's Bay Trail chips don't. Therefore, any tablet maker wishing to use a Bay Trail chip has to add separate chips for those functions, which would be an additional cost, hence the need for Intel's subsidy. Thus Intel is now paying the price for initially missing mobile; it has been locked out of the mobile arena, as far as mobile device system integration goes, and is having to crowbar its way back in.
The limitations of Tick-Tock
There's a second reason why Intel has struggled in mobile, which is perhaps more profound, and that's unlike in the PC era, chips are designed around devices and not the other way round. Back in the 90s, the most important parts of a PC were the processor, motherboard, and graphics card. The case was just packaging and the interfaces - called peripherals, appropriately enough - were commodities.
Fast forward to today, and the majority of personal computing devices (tablets, smartphones and desktops) are either made by vertically integrated companies, like Apple (AAPL) and Samsung, who also design their own chips, or they are from the likes of Nvidia and Amazon (AMZN), whose tablets and other devices nonetheless may have unique needs or particular functions, and in turn, specific requirements in terms of power, cost, performance and features for the chips which go into them.
That means it is no longer good enough for Intel to simply come up with chips that excel on a fairly generic set of performance metrics (but not have features like Wi-Fi or Bluetooth), nor can they set their product roadmap by the calendar and expect device manufactures to design around those chip designs - not when there's a host of highly versatile, fast moving and hungry Arm competitors who between them can cook up a veritable blizzard of chip designs at a moment's notice.
One also wonders whether the same competitive dynamics that have gone against Intel in mobile might soon apply to micro servers and after them, the real ones.
The problem with Foundry
If you can't beat them, join them. If Intel can't sell its own mobile chips to the likes of Apple or Samsung or (...practically anyone, actually), and if its contra-revenue gambit doesn't pay off, then one solution would be to make chips for others.
My main problem with this Foundry strategy is that in all industries, most of the value lies in IP/design and marketing, not manufacturing. Most of Coca-Cola's (KO) $178 billion market cap is simply the value of its brand.
And that's also the case in the tech industry: Apple makes vastly more money from designing and marketing its iDevices than Foxconn (GM:FXCOF) does making them. That's an extreme example of course, but there are plenty of other examples. Samsung started out as a component contract manufacturer - to Apple and others - but now is the most profitable tech company on earth through designing and selling its own devices. Qualcomm is currently the most valuable semiconductor company in the world largely through selling its IP.
Intel should know the value of marketing too. Its Intel Inside campaign was a huge success with consumers and helped it in winning the PC era battles with Advanced Micro Devices (AMD). Unless its foundry customers are willing to let Intel slap an Intel Inside sticker on their products - for free - Intel will start to disappear from consumers' consciousness.
I don't know whether Foundry is a bad business decision on Intel's part or worse, the correct one given Intel's weak competitive position, but either way, it would seem to be bad for Intel stockholders.