Originally Posted by shb-
This is nice and all, but i still cant see reason why it must be arm arch inside those simple internet browsing/word processing devices. I believe amd, intel or both of them will tune their low power x86 cores (bobcat/atom/whatever) soon enough (they are not so miserable at the moment afterall). Of course arm designs are better now, they have been paying attention to this LP market way longer then intel/amd, they arch are optimized for it (created with those LP ideas in mind). I really doubt intel will give up its market share without tough fight. Time will tell.
No one's saying it will be ARM necessarily that will dominate (although the odds are on their side as I'll explain shortly), but it's the ARM camp that will enter it first (I mean they are 100% pushing there already) and it's ARM camp that will improve it a lot faster at least in the first few years. The reason is Intel or AMD are not interested in these devices becoming more powerful.
Think about it, every device that is powerful enough for average joe will replace one of their desktop CPUs. Even if such a device has an Intel/AMD CPU, that CPU is going to be tiny and cheap (if they want a chance to compete with ARM devices at least), so no matter how you look at it, the mayority of their market shifted from an ASP of $150 per CPU + $100 mobo, to average of $25 per SoC. The company they are today is gone either way, no way to mantain their several dozen billion revenue. So Intel is going to fight the trend, the adoption of such devices as much as they can. The reason (all this is only my opinion obviously) is this, following the example prices from above:
- If the device is powered by an Intel SoC. They lost 1 desktop consumer (-$250) for one SoC consumer (+$25), that is, they lost $225 (compared to current market) and contributed to such devices becoming better, by simply being competitive. (because if the device uses Intel is safe to assume it is better than ARM)
- If device is ARM, they lost $250, but there's not as much competition as ARM is the only one really pushing for faster chips. Yeah, there's many people producing ARM based chips, but only one designing them and moving the tech forward. Big difference compared to 2 big companies/conglomerates fighting each other in designs and production.
The logical steps for Intel are clear for the first few years. Fight back sure, by creating SoCs, but do not compete entirely, do not create the kind of killer SoC that would take the crown yes, and the market share, but at the price of demostrating that those devices are (more) powerful. Use other methods to enter the market rather than using superior tech. This means that for a few years they will bleed desktop CPU market and will not gain too much market share on portable devices, but it's better than contributing to the trend and make the shift happen overnight. More years selling the $250 solution, even losing market, is better than being king in a less profitable (per sale) market, and everybody knows (even Intel), that they would not be king either, only another name in a long list of SoC manufacturers.
Like I said that's my opinion.