They do have good technology in 3D stacks with atom coming up, but whether that will materialize to dominating over the ARM ecosystem remains to be seen. If Intel can pull a big OEM like Apple towards their ATOM core, like they did in the desktop & laptop market on X86, they are back in the game in mobile computing. They do have a clear advantage in manufacturing. It will be interesting to watch......
Intel has come up to speed a lot in SoCs recently, something most other chip makers have been doing for a couple decades. It has got religion in low power--and will have the products to show it in about two more years, I think. But it has yet to buy in to the idea of a customer-driven business model. It has had the luxury of leading rather than following customers to date, and can continue that model as long as it wants to operate only in the PC sector.
I think it is choice above all. In the PC world, you practically have a monopoly (x86 and Windows) because it is a very low margin business and companies not only have to combat each other, they have to "fight" with anyone not afraid to put a card in a motherboard.
However look at the server side of things. Companies are paying very real money to develop their own Unix versions, to support Linux (almost all of the development is now funded) and not let single companies create monopolies (or at least monopolies that are not their own :) ).
The same goes for the mobile and embedded markets. OEMs need multiple sources. If they go the Intel route then they only have one choice, one foundry, one supply source. If Intel asks too much they have to bend, if Intel does not execute they have to offer inferior products, if Intel delays, they delay, if Intel cannot allocate enough capacity, then they have no products to sell. ARM is a magnificent solution to this problem and you can see how many companies are using this model of multiple sources. Samsung design its own SoCs but at the same time it offers devices with SoCs from nVidia, TI,... . The same goes for Nokia, LG, Sony Ericsson, whoever else. And I think that this is why even Apple who so dearly loves proprietary solutions, has chosen the ARM architecture because it knows that even it screws with its own chips (for any reason, even if it is capacity) then there could be a relatively easy backup solution. If it had used Power or x86, that would not certainly be the case.
Intel hates this choice model and unless AMD (or VIA) dont get their act together and offer solutions for smartphones or tablets, Intel cannot succeed. It needs others on its own court.
Could there be a more mundane explanation for Intel's emphasis on computers over smartphones? Intel is driven by Moore's "law". What began as an observation has become an obsession. Intel's focus is on doubling computing power in their chipset every two years. Smartphones require exquisite miniaturization and packing efficiency - a different focus.
I wonder if Intel while quick to drop an effort that is not panning out could benefit from getting in earlier on fewer projects? Does anyone know their track record? Wins versus misses? It would be interesting to know if they get it right more times than not and if their "average" was better or worse than the industry. Intel is a large player, but if ARM had got their act together and produced chips with much higher clocking and multi-core cpus earlier Intel would not have been the force it is and ARM would be in a better position.
Indeed it is strange to apply human attributes to a corporation, but I think "hubris" is the perfect word in this case. It's more likely the result of a pervasive corporate culture, a set of shared attitudes and beliefs, rather than control by one or a few individuals.
It is costly for them, that's all. They have their own internal 'venture capitalists' funding both research and development, and getting customers involved early means that things are more public.
They have a track record in pursuing new technology, sometimes it works and sometimes it doesn't, but their success doesn't correlate with other companies' experiences: for instance, no-one got bubble memory working either, but I think Intel made a Big Mistake abandoning RAM-based FPGAs especially when they were PCI-bus compliant. They would have had programmable SOCs years before anyone else.
(Yes, I got burned on Intel FPGAs but I didn't blame the whole company, it was a separate division [grrr])
Is funny how the behavior of a company can be described with adjetives that belong to a person. What does this mean? Will this mean that a company is controlled by only one person? Or a few? Or is it just the nature of a big organization? Is it that they arenīt aware of their place in the mobile computing world?
Interesting way to talk about the relation between Windows and Intel = Wintel.
I wonder what will the long term crop be for Intel if it keeps droping development of certain projects and affecting small development companies? Non good I bet.
Finding a market solution that enabled an x86-derived chip to address the mobile and/or TV markets without cutting profit margins for the server market was probably never an achievable goal.
Keeping the server profit margins for as long as the server market insisted on over-paying for x86 compatibility was definitely the right move.
Blog Make a Frequency Plan Tom Burke 17 comments When designing a printed circuit board, you should develop a frequency plan, something that can be easily overlooked. A frequency plan should be one of your first steps ...