You know I love a good Intel vs ARM fight and what’s going on in the mobile device market right now is nothing short of volcanic. ARM partners were all over CES and Intel demonstrated the first ever X86 smartphone and tablet on their Moorestown platform. Both companies know that a specialised platform is needed and both companies are enabling it. Who will win?
If you define ‘winning’ as ‘biggest market share’ then take a look at what ABI have to say. They think that ARM will overtake Intel in the ‘Ultra Mobile Devices’ space in 2013.
What does that mean?
ABI define a UMD as:
‘netbooks, MIDs, smartbooks and UMPCs’ although we had further clarification on the definition from Jeff Orr of ABI when we reported about 385 Million UMDs in 2014. Jeff Orr had this to say:
The 385m units in 2014 consists of Netbooks, MIDs, UMPCs and Mobile CE devices. Of those, about 158m are Netbooks. The Mobile CE devices include: mobile gaming consoles, PMPs, PNDs, digital cameras, digital camcorders, and eBook readers — all with mobile broadband connections.
So let’s get this right. Around 200 million mobile-broadband enabled devices in 2014 will have ARM inside. Around 80 million will be 3G ‘netbooks.’?
The numbers seem high but we’re talking about global markets and a wide range of devices. I have no doubt that 3G will penetrate the PND, PMP, Ebook, Mobile gaming and digital camera market. That’s huge!
One cross-reference is possible with Intel figures. They (and IDG) forecast 45 million netbook sales in 2012. If ABI and Intel are correct, ARM partners have a lot to gain in 2014!
Iphone data usage ‘a distraction.
Jeff comments in the press release:
There is a further important implication of this trend. Much is made in the media of certain mobile operators’ difficulties in providing sufficient network capacity to handle the growing demand for data. Users of devices such as the iPhone are often cited as the greatest source of that demand. Orr disagrees. “In fact, that’s a distraction, inch he says. “In general, laptops and netbooks with embedded or attached modems contribute a significantly greater amount of traffic to 3G networks than smartphones do. inch
On a global basis, that’s probably true.
Other thoughts from my side include the possible success of the Intel Moorestown and Medfield (and beyond) platforms. This is a real unknown at the moment.
RT @chippy: RT @umpcportal: ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7ABLJw
RT @chippy: RT @umpcportal: ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7ABLJw
RT @ARMCommunity:chippy: RT @umpcportal: ABI Research: #ARM-based cores will overtake x86 in Ultra-mobi Devices in 2013 http://bit.ly/7ABLJw
RT: @ARMCommunity: RT @chippy: RT @umpcportal: ABI Research: ARM uP Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7ABLJw
ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/8fXHrU #mobile
I’d believe it for UMPCs and netbooks when Microsoft releases their desktop OS to be both ARM and x86 compatible.
I had predicated a few months ago on another web site that Microsoft would
release a Win 7 for ARM when the shipments of ARM PC-type devices reached 50 million units a year.
This may still be possible, but the likelier scenario will be that Google
Android or Chrome or some version of Linux will be the OS for ARM devices.
This is based on the fact that Intel will scale the x86 architecture down to
where a low power version will be able to compete with ARM. The installed
base of x86 software and peripherals is simply too massive an advantage in
Intel’s favor. So there will be no need for Microsoft to port Win 7 (and
follow-on OSs) to ARM.
The only way ARM can compete is if it can convince application developers for
Windows to port their applications to a Linux based OS. Alternatively, ARM
can scale their processors so one or more cores can be devoted to hardware
emulation of x86. Hardware emulation isn’t as bad as it sounds (Apple was
able to offer software emulation of x86 when transitioning from PowerPC
to Intel architecture).
Intel is unmatched in process technology (at least 1 generation ahead of
everyone else, even IBM). Its microprocessor design team in Israel is
the equal of anyone else in the world. Worldwide, Intel boasts of about
12 fabs which can crank out product in tremendous volume. Intel also has
a lot of cash on hand. These resources mean that a paranoid Intel can
fund a massive effort to battle ARM (if Intel hasn’t already started
doing so).
The settlement with AMD has enabled Intel to concentrate on ARM.
@Chippy: For the forecast of UMD processor marketshare, mobile CE devices are not counted. The processor splits between x86 and ARM instruction set are comparing Netbooks, MIDs, Tablets, Smartbooks and UMPCs (and any other form-factor of mobile computer that might emerge between the laptop and smartphone).
@aplio: Don’t forget we’re talking about “always connected” devices here, especially with the ARM-based devices. When is the last time you put a DVD or an SD card into a device to watch a video? Compared to when you last went to a URL for video? Porting of x86 Windows applications is less relevant if the apps can be delivered in the cloud.
Thanks for the clarification Jeff.
Any chance you can give us the total number of umd devices (not including ce devices) for 2013?
Why exclude ce devices?
Steve
I’d like to know how soon we will get sub-$200 Chrome OS ARM Powered laptops with 20 hour battery time on a small battery, Pixel Qi screens for e-reader tablet modes and built-in free wireless broadband using White Spaces. I think that Google can make this happen this year already.
Intel cannot compete below the 4 inch screen size. ARM is too efficient, and got it all covered. Above 10 inch, I would go with Intel all the way. The good low priced devices are stuck at 10 inch.
The trouble is my perfect MID would have a 6 inch screen and no keyboard. Most available products in this range in America have Intel though. 3G can be added to those that lack it.
ARM and specifically Snapdragon are making announcements and release of phones like every week. The speed of change is with ARM. I’ve even considered tapeing a MIFI to the back of an Archos and calling it good.
Will Intel {and it’s partners} sit at the 10 inch netbook this year, and wait for ARM to jump up to the same size screen? Or will they start attacking ARM on size and style?
If you could spend $400, fit in a pocket, take phone calls, get the internet, be able to read it, and add common programs, what would you buy? What would everyone else buy?
I am so tired of built in obsolesces. Why can’t someone get to a MID size with their latest and greatest?
@aplio: Don’t forget we’re talking about “always connected†devices here, especially with the ARM-based devices. When is the last time you put a DVD or an SD card into a device to watch a video? Compared to when you last went to a URL for video? Porting of x86 Windows applications is less relevant if the apps can be delivered in the cloud.
……..
Intel cannot compete below the 4 inch screen size. ARM is too efficient, and got it all covered. Above 10 inch, I would go with Intel all the way. The good low priced devices are stuck at 10 inch.
……..
I will concede that for the time being, ARM will take the 4″ screen size and
below. And Intel will take the 8″ screen size and above.
On the OS side, Windows will rule on 9″ screens and bigger. Linux will battle it out with WIn CE on 4″ and smaller screens. Due to its expected
premium pricing (over $700), I don’t expect Apple to make as big a dent
here with its tablet as it did with iPhone. In surveys, people said they
were not willing to pay more than $700 for an Apple tablet.
So the battleground will be the 5″ to 7″ screens. I beg to disagree that
devices need to be always connected, and that the cloud will
host all the apps. For one thing, cellular carriers have a 5 GB monthly
data transfer cap, and mobile broadband rates are likely to go up (Verizon
Wireless recently raised its mobile broadband charges while lowering voice
charges).
The problem is also that no matter which CPU is used,
the sheer size of the display and other power-hungry elements in the device
will necessitate an intermittent connection mode, similar to that on winshield
wipers, as well as individual hardware and software switches for each radio/tuner (cellular, Bluetooth, WiFi, TV, etc).
I still stand by the importance of a sizeable application base. There are many problems with having one set of applications for the desktop and another set of applications for the mobile device. These problems discourage growth
in the number of users beyond the early adooters. No matter what the device, people will look to use Office type applications on it, plus Palm style personal information managers.
If fuel cells take off, Intel will muscle its way to market share leadership.
Devices with sufficiently large storage and screens will be used as entertainment devices, GPS receivers, TV sets, terrestrial and satellite radios, digital video recorders, podcast receivers, ebook readers, etc. Watching a full length movie is uncomfortable on a 4″ device. People want to tote as few devices as possible. The challenge will be to come up with an
uncluttered and extremely usable user interface. (This is where the
Apple tablet could make its greatest impact.)
RT @ARMCommunity: ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7ABLJw #fb
ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7PyGt0
ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 http://bit.ly/7PyGt0
Wont be long before ARM enters desktop market also. Just check out Marvell EBOX prototype and you can clearly see where ARM ecosystem is heading.
Unless ARM can emulate the x86 instruction set, it will be a
niche player. Businesses and the mass market are very hard to
change. If you don’t believe this, consider the following:
– within 1 year of netbooks being introduced, the majority
of them had a Windows operating system.
– COBOL applications are still running on mainframes. The
cost to migrate is simply too expensive.
– many netbooks sold in third world countries, which initially
came with Linux, are reformatted and Windows installed on them.
For all its technical excellence, Linux suffers from the lack
of breadth of applications, which I call the “one of each”
problem. This problem hounds any platform which lacks such
application breadth (such as the Mac).
Basically, the “one of each” dilemma is: a non-Windows platform
might have excellent software, but there is only one software
package or a handful in many categories. In other words, there is no choice.
So, someone looking for other features, a lower cost software
package, etc., does not have choice.
This lack of choice drives away consumers, handicapping the
platform’s acceptance.
“Unless ARM can emulate the x86 instruction set, it will be a
niche player. Businesses and the mass market are very hard to
change. If you don’t believe this, consider the following:”
Problem for Intel is that x86 instruction set will become niche and not another way around. ;D
Most 3rd world countries will become 1st world countries in next 50 years and many 1st world countries will become 2nd tier countries. Many folks in 3rd world countries don’t even know how PC looks like let alone UMPC’s, netbooks or smartbooks or what OS is.. it will be all around cost competitiveness and i doubt Intel will survive in the market where ARM licensees know how to make profit on low profit processors.
There are no 8GB x86 netbooks for $199 or less. Intel and Microsoft basically forbids the implementation of cheaper netbooks, where if a manufacturer tries to use Intel Atom processor to make a 8GB flash memory based 8.9″ or 7″ $199 Linux netbook, Intel suddently does not offer “advantegeous pricing” for the processor, which thus makes the cheap price point impossible to be reached.
In the same way, Microsoft only applies rebates for Windows XP netbook edition if the manufacturer basically promises to only ship Windows on all devices. A breach of that promise means the manufacturer would have to pay full price of $80 per Windows licence which makes it thus impossible to sell cheap netbooks.
Thus ARM is an alternative to Intel and Microsoft, not only in terms of being a technical alternative, that costs less and consumes less power, most importantly it is a competing architecture which dozens of providers of the processor can compete to provide differentiation, to provide lower cost, to provide lower power consumption.
You don’t need to “emulate” the x86 instruction set. You can run Citrix or some other hypervisor using Virtualization and you could run any X86 application that you would like just as smoothly as natively. In fact the experience would definitely be faster and better than on an Intel Atom processor.
Finally, people are using the browser only for everything now. Native x86 apps are not needed anymore. Everyone can notice the trend shift.
ABI Research: ARM-based Processors Will Overtake x86 in Ultra-mobile Devices in 2013 – http://shar.es/aSoZQ
that’s great and all, only a pitty ARM is not supported by Windows.