categories

HOT TOPICS

Subscribe to our Feed

Apple’s Component Strategy Shifts

Posted on Wednesday, Apr 23rd 2008

Apple just bought a 150-person chip company, P. A. Semi, to get its own low-power microprocessor design capabilities in-house.

Dean Takahashi does a good analysis over at VentureBeat, saying that this must be for some new product, not the iPhone.

Om Malik chimes in to say that this would pose a problem for Intel’s Atom chip. More on the Intel woes around this decision at Ars Technica.

As you know, I have been covering the convergence device component ecosystem at quite a bit of depth, and my primary reaction to this announcement is a very mixed feeling.

I am not concerned about Intel so much, as I am concerned about Apple itself, and its ability to manage this additional level of complexity that one more layer of vertical integration brings on.

Yes, I see the advantages. A new ultra low-power, ultra-high performance processor is needed to sustain the needs of the new category of devices that will replace laptops, integrate cell-phones, but will, at the same time, be powerful miniature multi-core computers. Apple wants to own the brains of this product, and not have to share.

Very well, if all goes well. But Apple has introduced great risk in its execution abilities.

Meanwhile, Intel should go talk to HP and DELL, both of whom need competing products in this category, and better get their acts together.

Hacker News
() Comments

Featured Videos

Comments

What makes Intel a non player in this game?

I see, in Intel a fierce competitor who will not let Apple go unchallenged.

And I am wondering, if acquiring a new Si company is indeed a better solution as with Atom series of processors, one end feature will be taken care of power consumption.
Yes there are other issues, in developing convergence chips, but I guess with Intel’s 32 nm process, it can surely pack a punch.

What do you think?

Soham Das Wednesday, April 23, 2008 at 12:05 PM PT

Where did you get that Intel is a non-player in this game?

I just said that they should align with their traditional partners and work on competing products.

Of course they can compete. We’re talking multi-core, low-power micro-processors … of course they can compete.

Sramana Mitra Wednesday, April 23, 2008 at 12:10 PM PT

it’s more than just owning the brains of the product IMO

It’s a long term strategy IMO

http://tinyurl.com/5gvyed

Sach Wednesday, April 23, 2008 at 1:36 PM PT

It is interesting that Jobs wants to control as much of his product as possible, yet he is able to get media companies to cede control (distribution) of their assets. Of course, the media companies were asking for it because of their clulessness of new media.

Dasher Wednesday, April 23, 2008 at 2:32 PM PT

It is still not clear to me why Apple had to buy this company as opposed to buying the chip they needed for MID? The latter would give them more flexibility in the future, and would be much cheaper… It really seems they have a greater plan on hand we are not aware of…

Denis Wednesday, April 23, 2008 at 2:46 PM PT

I just wrote a much longer and thought through piece on this announcement which will be my Forbes column for this week.

Stay tuned for it on Friday morning!

Sramana Mitra Wednesday, April 23, 2008 at 7:03 PM PT

Apple is not going back to PowerPC for their mainstream machines for a decade, if ever. You can only make your developers rewrite their apps for a new architecture so often. They hate you but they’ll do it once every ten years. Besides Apple’s had too much success with Intel to kick ‘em to the curb now. it’s more that Atom won’t be ready for phones for a couple of years. Even with the upcoming 32nm version.

Anyway, all iPods and iPhones use an ARM variant NOW. Apple is hyper about battery life and device size NOW. Apple needs to make ever more powerful but ultra efficient devices NOW to stay ahead of the competition.

Now, let’s think here, PA Semi is run by an ARM genius. Hmm… PA Semi is also fantastically successful at making high performance chips ultra efficient. Hmm… It’s a company that took an architecture that IBM and Motorola BOTH had serious trouble with and did what they couldn’t or wouldn’t – make it fast, but run cool. Hmm… It’s also been reported that Apple doesn’t want their current chips or their roadmap. Hmm…

So, what chip family does Apple use that this company might be SLIGHTLY familiar with (as in helped design), has the expertise and IP to make said chip more powerful and more efficient than it is and that might give Apple an edge in a hyper competitive market? It all really points to a custom ARM processor in my mind. At least for the near term. Apple wants ultra small sleek products with as little compromise as technically possible. Right now that means ARM in the tiny device space.

Atom probably won’t surpass ARM in the phone/ipod realm for several years. Especially if PA Semi can work their magic on ARM with any success making it even better than it already is. If and when Atom does surpass ARM, Apple is already able to slap it’s software on it so it’s no big deal for them. This isn’t a poke in the eye for Intel at all. In fact I’d bet we see it in other products from Apple within a year. Atom will get there but not fast enough for Apple to use anytime soon in iPods/Phones.

For less than $300 million Apple picks up a company that may be critical in keeping/extending their ability to compete in a multi-BILLION dollar market creating smaller, more powerful, longer running ‘i’ devices. I like their odds.

Ken Wednesday, April 23, 2008 at 9:14 PM PT

[...] analysis over at VentureBeat, saying that this must be for some new product, not the iPhone. Om Mahttp://sramanamitra.com/2008/04/23/apples-component-strategy-shifts/Crooks set sights on cellphone, laptop data Windsor StarEDMONTON – Like life, technology seems to go [...]

hp computers Thursday, April 24, 2008 at 3:31 AM PT

Sramana,

My thoughts exactly.

Except I’m not very impressed by the power numbers and the performance spec of PA Semi. I’m reviewing their white papers and shall blog over the weekend about this. (eg. Snapdragon has better numbers. XScale from Marvel has great numbers too)

I just feel power consumption is now a strategic advantage for any company in the ultra-mobile low power space. It’s risky to rely on other component providers to deliver on the power numbers. And what prevents them in selling the same solution to other providers?

Apple is building a competitive wide moat. This is part of that strategy.

Stay tuned.

Thanks,
-Sajal

EDIT: PA Semi website is down!

Sajal Thursday, April 24, 2008 at 3:37 AM PT

They’re buying capability, not products, Sajal.

Sramana Mitra Thursday, April 24, 2008 at 9:35 AM PT

No one can compete with Intel without fabs.

Jerome Friday, April 25, 2008 at 5:13 AM PT

[...] Apple’s recent acquisition of low power processor start-up PA Semi has caused a ripple in the semiconductor industry. Read my Forbes article on the acquisition in which I peek into Steve Jobs’ mind. I also say in the article that if Apple succeeds in developing a breakthrough, the industry will scramble towards vertical integration. [...]

Does Apple’s Move Impact ARM? - Sramana Mitra on Strategy Wednesday, April 30, 2008 at 8:41 AM PT
`