AMD Could Add ARM Faster Than Intel | Hacker News Reader

AMD Could Add ARM Faster Than Intel

blogs.forbes.com

19 pointsOsiris8 comments

A classic case of Innovator's Dilemma (Clayton M. Christensen) where ARM's low power chips are slowly attacking x86 from below. Intel/AMD tried to move upstream (Itanium) but found there were less customers who were looking for high power silicon, instead the rest of the industry is pivoting towards power per watt. Huge disruptions for the incumbents.

I'm not really sure the premise of the article, that Intel and AMD would actually want to add an ARM side to their business, is actually that true.

ARM chips have to be sold at very low cost - no one pays a premium for a particular brand ARM chip, but because of the recent market changes the R&D costs of ARM have gone through the roof and people like Apple and Nvidia are spending a small fortune in pushing the technology forward.

I think both Intel and AMD will be waiting for the market to develop further, then they can choose to enter the market if they need to.

>I'm not really sure the premise of the article, that Intel and AMD would actually want to add an ARM side to their business, is actually that true.

To me the article was saying that although there may be pressure to develop ARM chips, they don't want to:

"It’s not that Intel can’t switch to ARM; it’s that it most likely won’t."

"Intel has ... a philosophical predisposition not to back an alternative."

I thought it was a pretty good analysis.

Intel actually got out of the ARM business in 2006, pretty much exactly 1 year before release of the first iPhone: http://en.wikipedia.org/wiki/XScale#Sale_of_PXA_processor_li...

They apparently still hold an ARM license, but that may well just be so they can keep an eye on the competition.

I'm going to go out on a limb here and claim that Intel and AMD could probably produce an x86 chip that's comparable in efficiency to the ARM offerings by taking out some of the legacy cruft and substantially reducing clock speeds. After all, the Atom (and Llano) only consume around 10 watts and are still a few times faster than the current tablet CPUs. They may well enter the market at some point, but right now it's probably more profitable to carry on with their existing business.

People no longer care that there is "Intel Inside" when it comes to cell phones and tablets, the fastest growing markets. Sure, the Intel/AMD chip will support all of the x86 baggage, but what phone or tablet has legacy code to support?

right now they don't care, but I could see there being an 'intel powered iPad' marketed in future years - a good way to market the device as 'speedy' based on years of intel's own marketing.

So after years of buying the cheapest computers (cell phones and tablets), people will want to pay more for a name brand processor?

Not those same people. Right now a tablet is either `the cheapest computer' with appeal to low-end market or `the ultraportable' with appeal to some of the middle- and high-end market.

In a few years a tablet could well become `the powerful, ergonomic and ultraportable computer', with appeal to whole middle- and high-end market. That's the unwritten assumption behind the article: that technical progress will allow tablets to go the PC way -- decades of domination on wide market, rather than the netbook way -- cornered to low-end market, little market dynamics after few years.

Of course, Apple knows how to sell upgraded hardware to the same people every year or two -- but you can't have the whole market work that way.

Reply on news.ycombinator.com