I’ll Bet an ARM and a Leg That x86 Won’t Power Google’s Chrome OS Netbook


We’re still a good year away from the first Chrome OS netbooks that may come directly from Google (s goog), but more interesting to me is which processor will power them. Today Mike Arrington at TechCrunch says he’s willing to “bet one of our writers’ right hands that it’s ARM.” I don’t think there’s any need to sacrifice a hand, nor an ARM or a leg — my gut says that Google-designed device will go ARM all the way for a few reasons.

We already know that NVIDIA (s nvda) is working with Google closely on a device running Chrome OS. The company shared that news with us back in September and specifically mentioned the Tegra line of processors. Tegra — an ARM processor — is what powers Microsoft’s ZuneHD (s msft), and in my hands-on usage with the portable media player, it does a stellar job with the 480 x 272 display of the Zune HD. And that same device can pipe high-definition 720p video that looks fantastic. Clearly, the current Tegra is capable for digital audio and video and it’s not even the most current or powerful iteration of ARM out there. Inside the ZuneHD is a Tegra APX 2600 chip, which is based on the ARM11 architecture or ARM version 6. New handsets today are already running on the more powerful Cortex A8 architecture with the ARM version 7 instruction set.

And I doubt NVIDIA is sitting still with Tegra. I fully expect that at the Consumer Electronics Show, we’ll be hearing details about the next generation Tegra and maybe even get a peek at some prototypes. I actually have a pre-brief call with NVIDIA tomorrow, so it could be that I get a glimmer of Tegra’s future this very week. (Note that the call is under embargo, so I may not have info to share immediately). For NVIDIA, getting a chip to power a Google-branded device has to be on the radar. The company is fighting tooth-and-nail to get its Ion graphics chip in today’s netbooks, but essentially is at a cost disadvantage — Intel charges less for an Atom CPU in netbook if Intel’s own graphics are paired with it. A netbook manufacturer willing to use NVIDIA’s Ion with Intel’s Atom pays more for the Atom processor and has to pass that cost along to consumers. Lower costs make it possible for Google to even give these devices for free as James predicts.

Want another indicator that Chrome OS could go ARM over Intel’s x86 (s intc)? Just look at where Google’s consumer presence is felt in terms of mobile devices: Android. Possibly as early as 2011 we’ll see more Android handsets sold than iPhones. And every one of those Android devices is powered by an ARM processor. OK, maybe not every single one of them, but my Android on x86 effort is just an experiment. While the Chrome OS is very different than Android, both are based on efficient Linux kernels that don’t need the full power of an x86 chip. And the perfect storm of power vs efficiency might hitting by the end of 2010.

Want more netbook analysis? Subscribe to GigaOm Pro!

Until recent years, computing was all about processor power and clock cycle speeds. We wanted faster processing, richer third-party client apps and the ability to multitask til the cows came home. x86 computing is great for that. But it comes at a cost in terms of battery life. If you can only multitask for two hours before your battery burns out, what good is it in this mobile day and age? And so a shift towards efficient computing took place. The devices might be slower in this case, but they can run for six or eight hours. That’s the main reason I embraced netbooks when they arrived in late 2007 — they can handle the tasks I do today for nearly eight hours on a single charge. Let’s take the paradigm further though.

Google’s Chrome OS is all about a browser and nothing but a browser. Multi-tasking is accomplished within tabs, but still in the browser. The web is your client and all of your apps are in the cloud. Will you really need the extra “oomph” of an x86 processor on this type of device at the end of 2010? For most web tasks, probably not. In fact, if you’re using the browser on a Palm Pre (s palm), Apple iPhone 3GS (s aapl), Motorola Droid (s mot) or Nokia N900, ARM is already powering your browser. And for streaming video with either Adobe’s Flash or inline with HTML5, a next-generation ARM processor ought to work well enough by then. I’ve watched inline YouTube vids on the evaluation Nokia N900 handset and they play well. And so does using the browser for Google Docs. Simply put, the power efficiency of the ARM architecture beats that of the Intel Atom. And the ARM platform’s processing power has matured beyond the clunky PDAs and smartphones of 2003 — by this time next year, it ought to be capable enough for a web-based device that needs to run all day on a single charge.

There’s another aspect in all this that’s getting overlooked — Google is essentially playing in the operating system space with Chrome OS. Obviously the major players here are Microsoft… and Intel. It’s certainly no secret that Microsoft is a Google competitor, but don’t think that Intel isn’t in the firing line here. It placed a stake in the ground with its backing and development of Moblin — which currently is specific to Intel’s own Atom. My first look at Moblin on a netbook show it to be very capable, but lightweight environment geared specifically for smartbooks, netbooks and handhelds. If Google is getting in this space, do you think they would put an x86 chip inside that could make it easy to swap the Chrome OS for Moblin or Windows? I certainly wouldn’t if I had plans to become a market leader in this segment.

Now Google has already said that Chrome OS will be supported on both ARM and x86 platforms. I’ve already taken advantage of that by making Chrome OS the primary OS on my x86 netbook (see a video demo below). I’m not suggesting that you won’t see Chrome OS devices running on Intel Atom about 10 to 12 months from now. But from where I stand, ARM is very likely to power a Google-designed and possibly Google-sold Chrome OS device next year. The chip is cheaper and more power efficient that Intel’s Atom, which keeps costs lower and allows for longer run-times on a single charge. And while an x86 processor with the right graphics capability might be better suited to watching 1080p video, are people going to use a web-based netbook with a small screen for that? I doubt it. The use case for this device is to embrace the web while on the go. Content creation is just as important to it as content consumption and I see no reason why an ARM processor of tomorrow couldn’t power a Chrome OS device.



One nice thing about Android is that Android applications run on a VM, so you don’t have to have applications compiled for Android-x86 vs. Android-ARM. So if Android is the OS platform, the CPU is a matter of price and performance, not Intel x86 compatibility.

JavaScript: Apple and Google have spent quite a bit of effort in making JavaScript fast, their JavaScript engines are more than an order of magnitude faster than what was available in IE and Firefox before.


You can play 1080p with ARM no problem as well. Especially with the ARM Cortex A9 architecture which is sampling now and to be in commercial products by the time Chrome OS comes out. Texas Instruments, Nvidia, Qualcomm, Freescale, I think all of the ARM processor makers have or have announced 1080p playback support.

Also, playing back 720p or 1080p video on an ARM Powered Chrome OS laptop consumes less power than using an intel powered laptop.

I filmed an $80 Android laptop: http://armdevices.net/2009/11/12/80-android-laptop-menq-easypc-e790/ Now sold for $99 by the company Cherrypal. Consider simply the latest ARM Cortex processor in that, a Pixel Qi screen, optimized Chrome browser in Chrome OS for ARM, you’ve got a $100 ARM Powered Laptop, with 20-40 hour battery life and weeks of active standby. Consider also that the actual cost of a 15″, 13.3″ or 12.1″ screen is most likely not much more than $10 or $20 more than a 10″ screen. So that would be the cost difference using ARM for using larger screens and keyboards on the cheap laptops.

ARM has to show that it is ample powerful enough for fast and smooth processing of optimize Javascripts and Flash in the Chrome browser. This is what I ARM will be able to show as soon as optimized Chromium OS for ARM is being showcased.


I can see Google use a ARM processor for their Netbook or MID. Like Kevin said and I have been using a lot of the Google Sync, etc… which means that if you have a Netbook or some kind of MID running a ARM processor that would be just fine to use for that situation. You will be able to get that great battery life, light weight. For example, I have the Archos 5 which has a small screen but I am able to do mostly all my work stuff with it, so if you had a 10″ or 11″ Netbook with a ARM processor you could go all day and be able to do your basic work with it. For the most part anyone using a ARM Netbook or MID can do most of what they need to do with it, and have great battery life. Hopefully Kevin can share some more info tomorrow after talking with Nvidia about what’s going to be coming next year.


I hope what you speculate is true so long as Google gets Adobe to make a GOOD Flash port and also Java would be a very big plus.


You may not know that the CEO of Intel is on the board of Google. That may or may not mean anything but is certainly relevant…


Can’t wait to see how great the video will look w/ hardware optimized VP6 video. On2+Google = video bliss.


another HUGE advantage I think you left out is that a complete ARM based solution will allow for several days of active-standby (still get ur emails, tweets, etc) something x86 just can’t do.

ChromeOS not a true cloud OS thin client, computation is still done on the local machine. A Google book will not be a dummy-terminal used just for display while offloading all work to the server.

I wonder if Intel is kicking its own ass for selling its ARM division off to Marvell?


Does Javascript not work properly on an ARM? It perhaps has its best incarnation in Mobile Safari and WebKit.

ARM is the future, and probably has been since at least 2000. Hell, I recommended my future mother-in-law look into ARM stock in 1995. The scenario is perfect for establishing a heterogeneous ecosystem of processors that have to work on the same tasks. And, it fits perfectly with Google’s vision of the web being the primary processing component.

2010 is going to be quite interesting!


Kevin, in my opinion the achilles heel in your article is: JavaScript.

As we move more and more into the cloud and onto browser only environments our dependence on JavaScript is going to sky rocket. The omph as you put it that was needed out of the x86 chips is going to move into x64 performance in the JavaScript engine.

Intel apparently has chosen to include x64 in their next netbook chip so I don’t think things are going to be as cut and dry as you’re predicting.

Kevin C. Tofel

Great point on the JavaScript dependency Scotty, but are you suggesting we’re going to need 64-bit computing for a lightweight, inexpensive browsing device? You might be right, but it seems like overkill to me for this type of device.

I also don’t think that Chrome OS netbooks will be in an either/or situation, meaning we will see both ARM and Intel powered devices. I’m simply suggesting that a true Google-branded and sold device would be ARM powered.


What exactly do you mean by 64 bit performance. The points I can think of are 64 bit virtual memory, 64 bit physical memory, extra registers, 64 bit computations.

I don’t think more than 4GB virtual memory is relevant in netbooks yet, even though physical memory may be in a couple of years. The extra registers argument does not hold because 32 bit x86 was just register starved at 8 registers (ARM has 15), so 32 bit ARM does not have as much register spilling. As far as the 64 bit computations, I think Atom performs two 32 bit operations to get a 64 bit result making atom slower when performing 64 bit operations.

Is there something else I’m missing on 64 bit performance?

Comments are closed.