"Intel Inside" smartphones in 2012
At an investor meeting reported by Reuters, Intel announced that their mobile chips will be driving some smartphones by early next year.
http://www.reuters.com/article/2011/...74G5DF20110517
Intel's has so far been stymied in it's mobile platform chipset plans, even with it's new LTE-compatible X-Gold baseband processors released this year. No announcement on which manufacturer(s) signed on with Intel.
http://www.reuters.com/article/2011/...74G5DF20110517
Intel's has so far been stymied in it's mobile platform chipset plans, even with it's new LTE-compatible X-Gold baseband processors released this year. No announcement on which manufacturer(s) signed on with Intel.
Comments
At an investor meeting reported by Reuters, Intel announced that their mobile chips will be driving some smartphones by early next year.
Wishful thinking lives at investor meetings!
At an investor meeting reported by Reuters, Intel announced that their mobile chips will be driving some smartphones by early next year.
http://www.reuters.com/article/2011/...74G5DF20110517
Intel's has so far been stymied in it's mobile platform chipset plans, even with it's new LTE-compatible X-Gold baseband processors released this year. No announcement on which manufacturer(s) signed on with Intel.
And they will be crap. Thanks but no thanks.
If they are willing to spend the money, they can be one node or half node in front of anyone else. That's if they want to. Leading up to their Infineon purchase, Intel primarily focused on personal computing systems from netbooks (Atom), PCs (Core), to servers (Xeon, Itanium).
If they can get the Infineon chipsets on the latest nodes, it should be competitive part in terms of power and performance.
For smartphone SoCs, Intel really hasn't tried. They are trying to shoehorn Atom into a smartphone package, but everyone should be skeptical that it could work. Atom is really a 3-10 Watt product. Intel really needs to design an x86 package for 0.5 W. I don't doubt that they can do that. It's just matter of them wanting to.
The economics in many ways are just wrong. These are $20 parts. Intel is used to selling $200 parts. They'd have to devote significant chunks of money to develop a 0.5W x86 SoC and it won't be for much gain to the bottom line.
ARM is the incumbent. x86 is ugly. There is no reason to go there unless there is a market reason to do so. But, ARM is the incumbent.
They've tried to do better than ARM in ARM's niche, multiple times. Did they actually do any soul-searching as to why previous tries failed?
I think the answer to that is no. Even when they were in the ARM business with StrongARM, they weren't thinking very clearly. The "PC" architecture of discrete CPU + discrete GPU + discrete mem controller + ... was much on their minds and still is dominating. After Netburst almost busted them, they essentially divested all interest in non-core business, namely PCs and servers.
Until they do that, they don't have much of a chance. I get this "not invented here" vibe. Sure, licensing fees are undesirable, but it's also undesirable to throw money at a problem they can't solve. Go in with a sub-par strategy and they're liable to get sub-par results.
Never underestimate their fab advantage. If Intel is willing, they can fab Apple A5 SoCs on their 32 nm node and make a pretty penny. But the current political climate at Intel is indeed x86, and x86 only; aside from EPIC which is at best a hobby.
It's a matter of will and vision, not technical deficiency.
Never underestimate their fab advantage. If Intel is willing, they can fab Apple A5 SoCs on their 32 nm node and make a pretty penny. But the current political climate at Intel is indeed x86, and x86 only; aside from EPIC which is at best a hobby.
It's a matter of will and vision, not technical deficiency.
I agree. Intel is tough to beat on their home turf. They do have plenty of technical expertise, but it's a home turf issue, a lot of their intellectual capital is chip fabbing and designing PC chips. Designing for power-sipping handheld devices is a niche they can't crack without a considerable change in mindset, and the politics is part of that.
To really beat ARM at their own game, they need to forget about x86. Software for servers, desktops and laptops don't apply well to this niche anyway. They might be able to compete in or against the ARM market, but only if they license the architecture or instruction set somehow, or make a whole new instruction set. ARM is almost 30 years old, maybe something can be learned about it and make a next generation instruction set. ARM got to where they were by not going nuts with spending transistors, Intel got where they were by going nuts with spending transistors, neither is bad in itself, each was targeting a different market. The same expertise outside its target market usually doesn't succeed without significant adjustments.